Gene editors grapple with standardization
As the first CRISPR clinical trials begin this year in the U.S. and Europe, a number of genome-editing questions remain unanswered. One of the big ones: how to best measure whether a therapy is altering its target gene without introducing errors in others. To address that and other issues in the field, genome-editing companies and U.S. regulators teamed up at a workshop at the National Institute of Standards & Technology in April to start working out common standards that will promote better scientific understanding and confidence in therapies.
The NIST Genome Editing Consortium is addressing fundamental questions about analytical measurements underlying gene editing. Among the most critical are “What DNA changes did you make when you did genome editing, and how do you know your measurements are giving you the right answer?” says Samantha Maragh, consortium organizer and leader of the genome-editing program at NIST.
The consortium has three goals: to generate a common lexicon for the field, to qualify assays and come up with reference materials that put everyone’s tests on a level playing field, and to suggest data and metadata that could be shared across experiments and trials.
Standards developed by the consortium will be voluntary and not part of U.S. Food & Drug Administration regulation. Nevertheless, FDA joined NIST to sponsor the workshop, which was intended to “clarify the field’s regulatory needs and challenges and inform the efforts” of the consortium, according to workshop organizers.
The project creates unlikely collaborators in companies fiercely competing to commercialize gene-editing technology. But “this presents an opportunity for us, precompetitively with others in the gene-editing space, to enact standards and move everything more readily to the clinic,” says T. J. Cradick, head of genome editing at Crispr Therapeutics.
To join, organizations must agree to privacy terms and ground rules about intellectual property and contribute at least $20,000 or equivalent support. Members to date include Caribou Biosciences; Corteva Agriscience, the agriscience arm of DowDuPont; Editas Medicine; Intellia Therapeutics; Integrated DNA Technologies; and St. Jude Children’s Research Hospital. More companies and federal agencies are in the process of joining. About half the 124 participants at the April planning workshop came from industry, a third from the federal government, and the rest from academia and elsewhere.
Many members are heavily invested in the CRISPR/Cas9 genome-editing tool. The nuclease enzyme Cas9, directed by a customized RNA guide, finds and cuts DNA where the guide binds. The body’s own repair mechanisms take over to repair the site. Given a revised DNA template, the body’s enzymes will fix the site to match the new template, allowing for genetic code revisions and even additions.
CRISPR is remarkably straightforward, inexpensive, and fast. Its use has mushroomed since 2012, and treatments are already progressing into human tests. The first wave of drugs will be highly personalized: Doctors withdraw blood or immune cells, edit them in a lab, and reintroduce them into the body. Crispr Therapeutics is starting the first CRISPR clinical trials in Europe, to treat a genetic blood disorder called β-thalassemia. And the University of Pennsylvania is recruiting patients for the first U.S. clinical trial of a CRISPR-based therapy for cancer. Additional trials are expected this year.
But CRISPR is not the only way to edit genomes, nor is it the first to the clinic. Older technologies rely on other DNA-cutting enzymes, such as zinc finger nucleases, meganucleases, and TALENs (transcription activator-like effector nucleases). These require customized proteins, and they take more time, expertise, and expense to develop. Founded in 1995, Sangamo Therapeutics, for example, now has therapies based on zinc finger nucleases in clinical trials for hemophilia, HIV, and lysosomal storage disorders. In one trial, a man with Hunter syndrome became the first person in the U.S. to have a gene edited inside his body—a Sangamo zinc finger nuclease therapy was delivered to his liver.
The consortium hopes to bring all the gene-editing techniques under one standards umbrella for easier analysis and comparison. The consortium’s official start date is June 1, so no formal decisions on lexicon, data, or assay standardization have yet been made.
Nevertheless, participants involved in planning have already drafted a list of gene-editing vocabulary that needs more careful definition—terms including “on-target” and “genome editing” itself. “If a technology is being employed by a very slim number of people and they are all to one purpose, then terms and concepts are narrow and well defined. But when the whole world is doing it, it fans out into a multitude of nuance and meaning,” says Thomas Barnes, senior vice president of innovative sciences at Intellia Therapeutics.
On the data front, the consortium is considering how to house shared genomic data, which metadata entries and file formats will be required, and whether it may also offer test data sets for qualifying computational tools.
Some of the trickiest work of the consortium, however, will tackle the measurements that show whether a therapy is editing its target gene correctly. Drug developers need to know how prone an editor is to errors, which could include cutting in the wrong place, snipping the right spot but repairing incorrectly, or refastening the wrong ends back together.
Not every error will cause a problem. But the wrong change to a vulnerable gene could lead to cancer. With genome-editing therapies now working inside patients’ bodies, “it’s really important to understand exactly what we’re doing to the genomes of living cells,” says Shengdar Tsai at St. Jude Children’s Research Hospital. “Even a low-frequency off-target effect could be relevant.”
Any natural double-strand break could do the same, of course. DNA in a normal, dividing cell breaks from natural causes about 50 times per cell cycle. X-rays, airplane travel, and extra sun exposure trigger additional breaks. That complicates calculating off-target edits. Researchers have to be careful they don’t overreport off-target cuts, Barnes says.
Additionally, off-target editing can vary widely depending on the cell type, the target DNA sequence, the life span of the editing nucleases, and in the case of CRISPR, the way in which the guide RNA binds to its target. Researchers have three independent sets of tools for pinpointing off-target edits: bioinformatics tools, assays in living cells, and biochemical assays. FDA prefers to see drug developers use at least two of those, Anna Kwilas, a biologist at FDA’s Center for Biologics Evaluation & Research, said at the workshop. “We are still uncomfortable with using any one method to identify the off-target effects.”
The consortium plans to help by creating shared reference materials. What form those materials may take is unclear, although one likely option is DNA designed to mimic on-target and off-target edits, NIST’s Maragh says. The consortium will also look at qualifying existing assays and set up strategies for evaluating the variability, repeatability, and reproducibility of assays. Currently, no single assay for wrong targets can be used in all cells or for all genome-editing techniques.
Meanwhile, new genome-editing tools are already in the wings. Some researchers, for example, are exploring nuclease-free methods of genome editing. One approach uses strings of peptide nucleic acids that can invade the double helix and trigger repair mechanisms without cutting the DNA at all. Separately, Novartis and others are investigating how the adeno-associated virus, currently used to add genes in traditional gene therapy, can now also be used to edit the genome.
Discussing the various techniques and then deciding how to create standards that work for all in genome editing is the point of coming together as a consortium. “Standardization across editing platforms will be challenging,” Precision BioSciences Chief Scientific Officer Derek Jantz told the workshop participants. “It doesn’t mean we shouldn’t try. As responsible scientists, it is incumbent upon us to police ourselves and really do the best we can.”