In fewer than five years, an important new gene-editing tool called CRISPR has radically changed the face and pace of biological research. The ability to quickly and cleanly remove and replace stretches of DNA has already inspired thousands of publications featuring the technique and led to the creation of a slew of biotech businesses hoping to capitalize on CRISPR.

CRISPR’s power to effortlessly target and tweak any piece of DNA seems limitless. Thomas Barnes is the chief scientific officer of the CRISPR-centered Intellia Therapeutics, whose founders include one of the inventors of CRISPR, Jennifer Doudna. He says there is “an ever-growing backlog of well-understood rare genetic conditions with little that people can do about them.” Barnes hopes CRISPR will change that.

By tackling genetic disease at its roots—mutations in the DNA—CRISPR could end thousands of ailments, Barnes and others believe. Multiple research groups and companies are hot on the tracks of unleashing CRISPR on sickle cell disease, hemophilia, cystic fibrosis, Duchenne muscular dystrophy, genetic forms of blindness, and, of course, cancer.

The hype is partly about CRISPR’s broad applicability, but CRISPR’s true promise is its potential for a one-and-done cure. Changing your DNA is a permanent fix. CRISPR—short for the “clustered regularly interspaced short palindromic repeats” in the bacterial immune system from which the technology was derived—is a two-part system: a customizable guide RNA and a protein called Cas9. The guide RNA directs Cas9 to any desired segment of DNA for editing. The Cas9 enzyme then cuts the DNA at that precise location, allowing for genes to be turned on or off or for the removal or insertion of DNA.

But editing the DNA of cells in a petri dish—or even curing a mouse of a disease—is one thing; making the hot new technology work in humans is a whole other challenge. Sneaking the gene-editing complex into human cells is no easy task.

It will take some fancy molecular maneuvering to get the bulky Cas9 protein and the negatively charged guide RNA into humans. To work its magic, the unwieldy gene-editing system first needs to get into the body, skirt past the immune system, and infiltrate its target tissue. From there, it must sneak across cell membranes, escape the acidic environment of the cell’s endosomes to find the nucleus, and then home in on the correct location on the DNA. In other words, CRISPR has a drug delivery problem.

The Cas9 enzyme and the guide RNA composing the CRISPR complex cannot be swallowed in pill form or simply injected into the bloodstream. And a one-size-fits-all package is unlikely to work for every condition, so researchers are eagerly testing old strategies and creating new ones to achieve a CRISPR cure.

David Liu of Harvard University says this delivery dilemma isn’t unusual for a new gene-editing technology, but “researchers now feel this incredible urgency and excitement because of the promise of using CRISPR for therapeutic applications.” Since its inception as a gene-editing tool in 2012, nearly 5,000 papers mentioning CRISPR have been published in PubMed. The CRISPR craze is reeling in polymer chemists, drug delivery designers, and bioengineers all helping move CRISPR from the lab bench to the doctor’s office.

“I’ve just never seen any field that progresses at this pace,” says Niren Murthy of the University of California, Berkeley, who cofounded a start-up called GenEdit, dedicated to CRISPR delivery, in February 2016. “There is nothing comparable to the competitiveness of the CRISPR field,” he says.

“From a delivery perspective, I am sure there will be all sorts of surprises,” says Kathryn Whitehead of Carnegie Mellon University. C&EN spoke with more than 30 academic and industry researchers about CRISPR’s delivery dilemma. Some expect success soon. Others are trying to temper expectations, pointing to the historically long time horizon for turning new technologies into treatments. But as Whitehead says, “If this is possible, everything changes.”

A CRISPR way to edit

CRISPR isn’t the first gene-editing technology promising to cure thousands of diseases. In fact, multiple studies of treatments developed using older technologies are now under way. Drug delivery guru Daniel Anderson of Massachusetts Institute of Technology points out that one of the most advanced programs is Sangamo Biosciences’ ongoing clinical trial to remove T cells from patients, edit their DNA to make them resistant to HIV, and reinject the modified cells. “So presumably, there are some genome-edited people walking around in California that they helped create,” Anderson says.

Sangamo is using an older gene-editing tool called zinc finger nucleases, a complex protein structure designed to bind and cleave a specific region of DNA. And doctors at the Great Ormond Street Hospital in London recently reported using a similar gene-editing technique called TALENs, which also recognizes and cuts precise DNA sequences, to engineer immune cells for a therapy that may have cured two infants of leukemia.

Both technologies have been around for longer than CRISPR has, with zinc-finger-based editing being in the works for more than two decades. They also both suffer from a limitation that has inhibited their widespread adoption: Each is a cumbersome protein complex that needs to be individually engineered for every new DNA target.

CRISPR, meanwhile, is easily adaptable. The Cas9 cutting protein remains the same for all applications, and to make a new edit, researchers need only to switch out the guide RNA. If the DNA sequence that needs editing is known, securing the complementary guide RNA is as easy as clicking “Order” from a supplier.

“When CRISPR came along, everyone knew what to do with it,” Intellia’s Barnes says. “People had been going around in a go-kart and you gave them a Ferrari, so away they go.”

Jacob Corn of UC Berkeley says cheap and easily customizable guide RNA empowers the “democratization of gene editing.” Corn’s lab is one of several using CRISPR to cure—at least in isolated cells and mice—sickle cell disease, where a single-letter DNA mutation stymies the oxygen-ferrying capacity of red blood cells.

Corn envisions a world where patient DNA testing is coupled to CRISPR’s customizability, and scientists can easily whip up a fix for problematic genetic mutations. “I think that, in the future, we’ll be able to tackle genetic diseases with the same speed we can diagnose them,” he says.

Moving into cells

Corn’s dream might not be far off, at least for blood disorders such as sickle cell disease. In that condition, stem cells collected from the blood or bone marrow could be removed from a patient, edited in the lab to correct the DNA typo—a process called “ex vivo” gene editing—and then reinjected to proliferate and make a patient healthy.

Editing cells harvested from a patient is relatively straightforward. Researchers commonly use electroporation, a technique that uses an electric pulse to momentarily create pores that allow the Cas9 protein and guide RNA complex to slip inside cells in a dish. This technique has the potential to address hematological disorders and is also being used to beef up immune cells to fight cancers such as leukemia.

Lloyd Klickstein, head of translational medicine for the new indications discovery unit at the Novartis Institutes for BioMedical Research, says, “Thus far, the ex vivo technologies are what’s been done, and that’s what most of the companies are looking to do first,” Novartis included.

The concept looks promising on paper, but no one knows how well it will work in humans. Chinese researchers at Sichuan University claimed to be the first to do ex vivo therapy with a handful of people with cancer last year, and University of Pennsylvania researchers are gearing up for a similar clinical trial in Philadelphia, San Francisco, and Houston this year.

Although researchers are excited about the potential to use CRISPR to create therapies from people’s own blood, immune, and stem cells, thousands more genetic conditions affect everything else. For those disorders, CRISPR needs to be delivered like more traditional medicines so it can work its wonders editing DNA inside the body. But the challenge of shuttling CRISPR directly to the diseased tissue, or “in vivo” gene editing, is so daunting it could stall CRISPR’s otherwise rapid advancement.

Déjà virus

The first in vivo CRISPR therapy to be tested in humans will likely borrow its delivery vehicle from the world of gene therapy, where hollowed shells of viruses are used to transport genes inside cells and then, in theory, permanently produce a therapeutic protein. Decades of gene therapy research has yielded a reasonably good carrier for genetic material, the adeno-associated virus (AAV). Compared with other viral vehicles, the immune system tends to ignore AAV, and the carrier is able to target specific cell types in the body.

Editas Medicine’s founding scientific adviser is one of CRISPR’s inventors, Feng Zhang of the Broad Institute. Editas is using AAV to deliver CRISPR in monkeys. In this study, CRISPR targets the genetic mutation that causes Leber congenital amaurosis 10, a rare form of progressive childhood blindness. The biotech firm plans to ask FDA for permission to start human studies of the treatment, which must be directly injected into the eye, by the end of this year. “We chose that disease because we felt that we could deliver our machinery there,” says Charles Albright, chief scientific officer of Editas.

Seokjoong Kim, research director at the South Korea-based gene-editing company ToolGen, is also conducting CRISPR experiments in mice for eye disorders, including age-related macular degeneration and diabetic retinopathy. ToolGen will also deliver CRISPR with AAV because it “is the most validated delivery tool clinically,” Kim says. He notes that AAV is already being used in gene therapy clinical trials for Parkinson’s disease, hemophilia, and vision disorders.

“We are building on decades of work in gene therapy,” Albright says. “We believe that patients and regulators and physicians will feel more comfortable using this method.” Using CRISPR in humans is enough of an unknown for Editas and ToolGen, and they believe the chances of success and drug approval are higher with an established delivery system such as AAV.

But AAV’s strength for gene therapy—perpetual production of a protein—is its drawback for gene editing. “One of the potential issues with AAV is that there is no good way to control the expression of Cas9,” says Mark Kay, director of the Program in Human Gene Therapy at Stanford University. Once inside cells, the DNA plasmid will continue producing the Cas9 enzyme indefinitely. Since CRISPR needs to make its edit only once, the longer Cas9 hangs out inside the cell, the greater the chance the enzyme will make unwanted cuts in a patient’s DNA, Kay says.

Those off-target cuts are frequent topics of concern among CRISPR scientists. Even though the tool is precise, there is no guarantee it will perform to perfection. That liability has driven many researchers to look for other ways of delivering CRISPR.

An RNA flashback

The challenge of commercializing CRISPR has an even closer cousin than gene therapy. Work by scientists in 1998 unexpectedly showed that double-stranded RNA molecules could suppress the translation of messenger RNA (mRNA) into protein. Known as RNA interference, or RNAi, the research garnered a Nobel Prize in 2006 and spurred the creation of start-ups aimed at turning this powerful method of silencing genes into therapies.

But RNA cannot be directly injected into the bloodstream, where it gets degraded and triggers an immune reaction. So scientists have spent the past decade figuring out how to get their molecules inside cells. Now, CRISPR researchers are hoping to borrow their most common delivery vehicle, the lipid nanoparticle.

To reuse lipid nanoparticles for CRISPR, the gene-editing system has to be packaged in a way that recapitulates the negatively charged RNA molecules used in RNAi. Instead of delivering Cas9 as a functional protein, many researchers are sticking the mRNA instructions to make Cas9 inside their nanoparticles and letting the cell produce the protein.

Anderson’s lab at MIT has been a center of thought for this research. Hao Yin, a postdoctoral researcher in Anderson’s lab, packaged Cas9 mRNA in lipid nanoparticles previously developed in the Anderson lab for shipping RNAi molecules across a cell’s lipid membrane. Yin then delivered that alongside guide RNAs packaged separately in AAV to fix broken genes in mice with liver disease. The upside of the method is minimal off-target cutting by Cas9. “The downside is that the efficiency is very low,” Yin says. Only about 6% of hepatocytes, or liver cells, were edited by CRISPR (Nat. Biotechnol. 2016, DOI: 10.1038/nbt.3471).

New and improved lipid nanoparticles are popping up that can deliver both Cas9 mRNA and the guide RNA in the same particle. Although that dual packaging would in theory improve editing efficiency, it also poses some logistical problems. The molecules used in RNAi are only about 20 nucleotides long. Guide RNAs used in CRISPR, on the other hand, are about 100 nucleotides long, and the mRNA encoding Cas9 is an unwieldy beast of 4,500 nucleotides. “So if you take the off-the-shelf lipid nanoparticle formulation and instead encapsulate the CRISPR system, it’s just not very good,” says Daniel J. Siegwart of the University of Texas Southwestern Medical Center.

Siegwart, who was previously a postdoctoral researcher in Anderson’s lab, became the first to successfully deliver that kind of dual packaging to mice in December. Lipid nanoparticles require several ingredients for ferrying RNA into cells, including positively charged lipids for binding the negatively charged RNA. Siegwart’s group synthesized zwitterionic amino lipids, ones containing both positive and negative charges, which help bind, stabilize, and release the mRNA as the particles cross into cells (Angew. Chem. Int. Ed. 2016, DOI: 10.1002/anie.201610209).

Lipid nanoparticles enter cells through a pinched-off cell membrane envelope called the endosome, and getting Cas9 mRNA out of that structure can be another limiting step. In January, Paul A. Wender and Robert M. Waymouth of Stanford University unveiled a polymer nanoparticle system to overcome this problem. Their particle acts like a “physical property chameleon,” Wender says, changing its form as it crosses the cell membrane and enters the endosome.

Wender and Waymouth’s system, called charge-altering releasable transporters, are made of initially positively charged oligo(α-amino ester) polymers that bind the negatively charged mRNA. Upon entering the endosome, where the pH becomes more acidic, positively charged amine molecules in the polymer become neutrally charged amides, which releases the mRNA into the cell (Proc. Natl. Acad. Sci. USA 2017, DOI: 10.1073/pnas.1614193114). Although their paper didn’t explicitly test the concept on Cas9 mRNA, that’s on the to-do list.

Protein power

Lipid nanoparticle innovation may be blossoming, and CRISPR developers are confident they can reach the clinic more quickly and safely than with RNAi, but that delivery vessel is by no means foolproof.

Rodger Novak, chief executive officer of CRISPR Therapeutics, whose founders include another of CRISPR’s co-inventors, Emmanuelle Charpentier, points out that “CRISPR has an advantage” over RNAi, which turns down protein production only temporarily and needs to be readministered periodically. Those repeat injections can cause liver toxicity, a side effect that has slowed down the initially rapid progress of RNAi companies. Although the technology is maturing, there are no approved RNAi drugs.

“The whole lipid nanoparticle field is a little bit weird,” Ross Wilson of UC Berkeley says. The literature is full of “success stories that are never followed up on; they just fizzle.”

Wilson is one of several researchers working on delivering Cas9 as a protein rather than as mRNA in a lipid nanoparticle or as DNA in a virus. Researchers call this form of CRISPR a ribonucleoprotein, which is the active form of the guide RNA hooked up to the Cas9 enzyme in a single, ready-to-go complex.

David Liu of Harvard University says delivering CRISPR in a virus gives “the least amount of control” because it manufactures the Cas9 protein indefinitely. If there are too many Cas9 enzymes in a cell, there is a greater chance that one of them may accidentally cut DNA in the wrong place. Directly delivering the protein gives “the most control” because lower levels of Cas9 in each cell means a lower risk of potentially dangerous off-target cutting, Liu says. His group developed cationic lipid nanoparticles for CRISPR ribonucleoprotein delivery (Nat. Biotechnol. 2015, DOI: 10.1038/nbt.3081).

Liu, along with Qiaobing Xu of Tufts University, also created lipid nanoparticles that are biodegradable inside cells. The system binds negatively charged CRISPR ribonucleoproteins initially, but releases them upon entering the chemically reducing environment of the cell (Proc. Natl. Acad. Sci. USA 2016, DOI: 10.1073/pnas.1520244113).

Wilson is looking to find a way to deliver CRISPR ribonucleoproteins without the hassle of lipid nanoparticles. To do that, he needs to make the ribonucleoprotein complex stable in the bloodstream, able to escape the cell’s endosome, and even able to home in on a particular tissue type. But there is a downside. “The immunogenicity of Cas9 could be a real issue,” Wilson says.

Other scientists are crafting even more exotic delivery systems for CRISPR, including a yarn ball-like structure called a DNA nanoclew developed by Chase Beisel and Zhen Gu of North Carolina State University. Their nanoclew uses repeated stretches of DNA complementary to the guide RNA wrapped up in a ball to deliver Cas9 protein to cells. (Angew. Chem. Int. Ed. 2015, DOI: 10.1002/anie.201506030).

Remaining challenges

Even as the field works out the delivery kinks, therapies are expected to soon reach people. Clinical trials using ex vivo gene editing in humans with CRISPR is anticipated to start in the U.S. this year, with in vivo gene editing likely in 2018 and 2019.

Casebia Therapeutics, a joint venture between Bayer and CRISPR Therapeutics, is making its commitment to the delivery challenge clear, with plans to hire a head of delivery. James Burns, CEO and president of Casebia, says, “There are some approaches that we can take now, but to really harness or achieve CRISPR’s full potential, we are going to have to invest in new delivery technologies.” Currently, most CRISPR-based companies are taking an agnostic, “whatever works” approach, testing both AAV and lipid nanoparticles for their first rounds of treatment.

Berkeley’s Corn points out another problem with CRISPR that many people conveniently gloss over. “We are really good at breaking sequences and not really good at fixing them,” he says. Some conditions can be cured using Cas9 to cut out a mutation or turn a gene off. But there are many more conditions where faulty DNA needs actual correcting. That requires a third component: a DNA template strand to tell the cell’s repair machinery how to fill in a cut made by Cas9.

“Delivering all three has been really challenging and it has not been demonstrated in in vivo systems with any lipid nanoparticles yet,” says Kunwoo Lee, who is now CEO of the start-up GenEdit that he founded in February 2016 shortly before finishing his Ph.D. in Murthy’s lab at Berkeley. GenEdit focuses on applications of CRISPR that will require a DNA template for repair.

Lee and his colleagues at GenEdit already have a few scientific studies under review, including one that uses gold nanoparticles as a core material to load the three components of the CRISPR system. They are also working on lipid and polymer nanoparticle systems, all designed to deliver CRISPR ribonucleoproteins. Although that strategy is promising for minimizing off-target cutting, it may also be the furthest away from being an injectable treatment in the clinic.

“There is simply no way that any particular delivery modality is going to provide the means to address all of those targets, so it really needs to be an all-of-the-above approach,” says Erik Sontheimer of the RNA Therapeutics Institute at the University of Massachusetts Medical School. And from the looks of it, the CRISPR companies are approaching it as such.

“I’ve heard many people say buckle up because there will be a trough of disillusionment that has to be traversed before it can become a clinical reality,” Sontheimer says. “But the potential payoff is so clear, that there will be enough staying power if and when that comes.” 

CORRECTION: This story was updated on Feb. 14, 2017, to correct the description of Casebia Therapeutics. It is a joint venture between Bayer and CRISPR Therapeutics rather than Novartis and CRISPR Therapeutics.