This past July saw a big leap forward for clinical application of CRISPR-based genome editing, with the launch of the Brilliance trial by Allergan and Editas Medicine. Previous clinical forays into genome editing have focused on manipulating isolated human cells in the laboratory, which are then transplanted back into patients. In contrast, Brilliance will be the first in vivo test of this technology in humans, with patients receiving direct injections of viral particles laden with genes encoding the CRISPR-Cas9 machinery to correct a retinal gene defect.
This progress comes amidst a general surge in progress with nucleic acid–based therapies. Indeed, many researchers working toward clinical translation of CRISPR believe their work has been buoyed by recent breakthroughs with RNA-based drugs such as Alnylam Pharmaceuticals’ Onpattro and Biogen’s Spinraza, as well as AveXis’ Zolgensma—the first gene therapy to win FDA approval. “All of these advances being made with virally and nonvirally delivered gene therapies are now being looked at for CRISPR therapy,” says Daniel Anderson, PhD, a researcher at the Massachusetts Institute of Technology and a co-founder of CRISPR Therapeutics.
Inventorying CRISPR cargo
But CRISPR also brings distinctive challenges, including a more complex set of parts that must be introduced simultaneously to achieve efficient editing. “The molecules are much larger than an antisense RNA,” says Krishanu Saha, PhD, associate professor of biomedical engineering, University of Wisconsin–Madison. “The Cas9 enzyme is a pretty large protein, and then it’s complexed with a guide RNA.”
These two elements suffice if the goal is to use CRISPR to introduce a targeted deletion in the genome, via a DNA repair mechanism known as nonhomologous end joining (NHEJ). For example, the therapy being tested in Brilliance, EDIT-101, uses NHEJ to delete a mutation that truncates an essential retinal protein.
However, other groups aim to exploit another mechanism known as homology-directed repair (HDR), which allows the targeted insertion or replacement of larger stretches of DNA. This requires the inclusion of yet another component, a “donor DNA” strand destined for genomic insertion.
“This is a big complicating factor,” says Ross Wilson, PhD, project scientist and principal investigator, Innovative Genomics Institute, California Institute for Quantitative Biosciences, University of California, Berkeley. “[Using HDR] can cause acute toxicity inside cells, which dislike having a lot of donor DNA in there.”
Once delivered, CRISPR quickly wears out its welcome. Although it can be extremely precise, the longer Cas9 lingers in the nucleus, the more likely it is to begin making unwanted, off-target cuts in the genome. There is also evidence that Cas9, being a bacterially derived enzyme, can elicit an immune response in some patients,1 creating other opportunities for side effects. Although the clinical impact of this immunity remains unclear, it provides further incentive to keep treatment transient. “The ideal scenario,” says Saha, “is that you deliver it well once, and then it’s gone.”
Looking beyond ex vivo routes
It is therefore unsurprising that the earliest clinical progress has been ex vivo, using isolated cells that can be individually manipulated and subjected to stringent quality control before reimplantation. In this context, researchers have access to a host of methods for efficient delivery into cells. For example, Beam Therapeutics is repairing mutations associated with diseases such as sickle-cell anemia using electroporation, in which an electric current is used to temporarily open pores in the cell membrane through which the CRISPR machinery can readily pass.
“It’s got clinical validation in multiple other scenarios, so there’s a good regulatory path forward,” says Manmohan Singh, PhD, Beam’s vice president of pharmaceutical sciences and delivery technologies. “And in most cases, you get 80–90% cell viability.” However, this method is clearly not suitable for manipulating large numbers of cells in living tissue, and most in vivo work is instead focused on the use of viral vectors or synthetic nanoparticles to achieve targeted delivery.
Deploying viral vehicles
Adeno-associated virus (AAV) has established itself as a vehicle of choice for gene therapy, and many in vivo CRISPR efforts are following suit, including Editas’ EDIT-101 program. AAV is nonpathogenic and has generally proven safe for delivering DNA to a variety of tissues. “You have these different AAV serotypes that you can use to go after certain cell types or organs,” says Singh, who also notes that the industry has well-established manufacturing strategies in place for the large-scale production of engineered viral particles.
Despite possessing several advantages, the viral capsid offers relatively cramped quarters—barely enough room for the genes encoding the Cas9 enzyme and guide RNA. Beam’s researchers have found this to be a particular challenge in the context of the company’s base-editing strategy, in which Cas9 is rendered catalytically inactive and coupled to a second enzyme known as a deaminase.
Base editing induces selective chemical modification of bases at the site targeted by the guide RNA—for example, converting adenine to guanine—without actually cutting the DNA. This approach enables precise correction of mutations with virtually no risk of the unwanted insertions or deletions that can occur during NHEJ. But it also produces a much bigger cargo to package, and the company therefore makes use of multiple vectors to get the job done.
“We split the base editor into two parts, and we load each of these AAVs with one half,” explains Singh. “Once the two proteins are encoded in vivo, they come together to form a fully functional base editor.”
There are other potential issues with AAV. “Some people have preexisting immunity to the most commonly used serotypes,” says Anderson. This could reduce the efficacy of therapy, and even nonimmune patients receiving multiple doses of AAV therapy could experience diminishing returns from treatment if they develop a response after their first administration—although most CRISPR researchers aspire to develop single-dose therapies.
Kunwoo Lee, PhD, CEO and co-founder of GenEdit, also points out that even though genetic material delivered by AAV does not get permanently incorporated into the genome, it can persist in the body for extended periods. “In some cases, it persists for weeks; in other cases, months,” says Lee, “and [the genetic material] keeps producing Cas9.” His team and others are therefore pursuing synthetic nanoparticle–based strategies as an alternative approach for delivery.
These nonviral approaches can offer greater control over how the CRISPR machinery is delivered, and thus how long these components linger in the cell. Some groups are opting to encapsulate ribonucleoprotein (RNP) complexes, where the Cas9 enzyme and guide RNA are preassembled to achieve rapid editing upon delivery without the need for transcription or translation. Others deliver these components as messenger RNA strands, which ensure the shortest possible duration of Cas9 activity. “Messenger RNA will give you a peak Cas9 exposure after six to eight hours, and then it just gets degraded and goes away,” says Anderson.
Alnylam’s Onpattro, the first RNA interference–based therapy to reach the market, uses lipid nanoparticles (LNPs) to deliver its therapy to liver cells, Beam is among several groups exploring a similar approach for CRISPR messenger RNAs. “This process is very scalable,” says Singh, “and there is enough understanding of lipid stability that you can really fine-tune your lipids to get enough stability to carry your cargo.”
Myriad polymer formulations are also available, and several groups have demonstrated successful in vivo genome editing in animal models. For example, Saha and his collaborator Shaoqin “Sarah” Gong, PhD, professor of biological engineering, University of Wisconsin–Madison, have identified a method for promoting self-assembly of polymeric coatings around Cas9 RNPs.2 The investigators have demonstrated that their delivery approach can achieve efficient genome editing in the retina and muscle of mice.
“We found a formulation that is as good as the best commercial agents, and that also has several advantages,” says Saha. For example, it can be freeze-dried, and it is predicted to be stable in the bloodstream. Saha and Gong are now developing their polymer formulation as a means of delivering CRISPR across the blood–brain barrier in neurological disorders,3 with funding from the Somatic Cell Genome Editing (SCGE) program—a $190 million National Institutes of Health effort to support clinical development of CRISPR.
GenEdit is working with an alternative polymer-based approach, based on research Lee conducted as a graduate student at UC Berkeley. The company is testing large libraries of polymer building blocks in various assays to identify the best match for a given drug delivery challenge. “We screen for the best one so that we understand why a particular polymer delivers better into a specific tissue, and with that rationale, we create a second-generation library and keep improving the delivery,” explains Lee.
One nice feature of polymeric nanoparticles is that they can easily be modified to achieve even more selective delivery—for example, Saha’s team has tailored its particles with chemicals that can target certain cells within the eye, or peptides that enhance cellular uptake. Less advantageously, however, each new polymer formulation requires careful characterization to ensure that it is both nontoxic and unlikely to elicit a host immune response—a risk that can be devilishly hard to predict in preclinical studies. “Even nonhuman primates don’t have an immune response that really recapitulates the human equivalent,” says Wilson.
Although most in vivo genome editing efforts choose either viral or nonviral delivery, Anderson’s group has shown that there are advantages in combining the two—particularly in the context of HDR-mediated gene repair. In this scenario, LNPs are used to deliver Cas9 messenger RNA while AAV particles carry the guide RNA and donor DNA sequences, thereby taking advantage of the strengths of both technologies.
In an initial demonstration, Anderson’s group was able to correct an enzymatic deficiency in >6% of mouse liver cells—a 15-fold improvement relative to previous HDR-based approaches.4 His group is now using SCGE funding to develop this approach as a treatment for cystic fibrosis, and Intellia Therapeutics has also made considerable headway with a similar combinatorial approach in in vivo studies with nonhuman primates.5
Approaching the last mile
These remain very early days for in vivo gene editing, as researchers strive to achieve levels of CRISPR efficiency and accuracy that can now be routinely achieved ex vivo. But Wilson believes that such progress will be essential for truly unlocking the clinical potential of this technology—not just in terms of the diseases that can be treated, but also in terms of the patients that can benefit from treatment.
“Even if a cell therapy is safe and potentially affordable, it’s still something you can do in only a few places on Earth,” Wilson says. “It’s not capable of being deployed to all these other places where there are millions and millions of people with life-threatening genetic diseases.”
1. Charlesworth CT, Deshpande PS, Dever DP, et al. Identification of preexisting adaptive immunity to Cas9 proteins in humans. Nat. Med. 2019; 25(2): 249–254.
2. Chen G, Abdeen AA, Wang Y, et al. A biodegradable nanocapsule delivers a Cas9 ribonucleoprotein complex for in vivo genome editing. Nat. Nanotechnol. 2019; 14(10): 974–980.
3. National Institutes of Health Project Information. Enabling nanoplatforms for targeted in vivo delivery of CRISPR/Cas9 ribonucleoproteins in the brain. Awardee organization: University of Wisconsin–Madison. Project leader: Gong, S. Project number: 1UG3NS111688-01.
4. Yin H, Song CQ, Dorkin JR, et al. Therapeutic genome editing by combined viral and non-viral delivery of CRISPR system components in vivo. Nat. Biotechnol. 2016; 34(3): 328–333.
5. Huang, H-R. CRISPR/Cas9-Mediated Targeted Insertion of Human F9 Achieves Therapeutic Circulating Protein Levels in Mice and Non-Human Primates. Presented at: 22nd Annual Meeting of the American Society of Gene and Cell Therapy. April 29, 2019; Washington, DC.