Not all news about CRISPR gene editing has been good. In 2017, a paper in Nature Methods claimed that the technique created hundreds of unintended mutations in mouse models, and that algorithms designed to detect those “off target effects” weren’t working. The news sent share prices of CRISPR-based companies plummeting, and created rumblings—even panic—among scientists and gene editing-watchers worldwide. How accurate was CRISPR-Cas9, after all?
But the paper turned out to be seriously flawed, and Nature ultimately retracted the study. Since, other work has shown that such off-target activity is rare, and not distinguished from natural mutations. One “response” study published in PLOS Genetics by the Wellcome Sanger Institute in 2018 used pedigree-matched animal model controls and whole genome sequencing to show that this technique allowed them to determine whether CRISPR-generated mutations were causing single nucleotide variations (SNVs) or off-target indels in embryos. This more precise examination allowed for the conclusion that such off-target mechanics were highly unusual, while previous studies used strain-matched animal models, which did not allow researchers to tell the difference between generation- or colony-related differences in genes and true new mutations. Still another paper, this one in Nature Biotechnology, showed the CRISPR off-target mutations were due more to using non-homology end-joining (NHEJ), versus homology-directed repairs (HDR).
Regardless of the 2017 paper’s flaws, it still underscored the need to use the right animal model to determine the precision and safety of CRISPR, especially as it enters biomedical fields as a key component of new drugs to treat diseases, vaccines to prevent them, and diagnostics to identify them.
Zebrafish, long a tool in development biology and in genetics, are emerging as a valuable animal model for CRISPR work, including the determination of off-target effects. In a paper published in 2018 in The Company of Biologists, PhD student Annekatrien Boel and her colleagues at Ghent University in Belgium did find that CRISPR “knock-ins” introduced a number of mutated patterns of repair templates in Zebrafish. This wasn’t supposed to happen, because their alterations used the homology directed repair (HDR) method, which scientists assumed didn’t introduce these off effects. This is especially important in addressing diseases caused by specific point mutations, which arise from an altered single nucleotide pair.
“It’s important to recognize the manifestation of erroneous repair during HDR, which was until now generally considered to be error-free, and to apply the appropriate techniques to analyze the efficiency and correct of editing,” Boel said in an interview with the journal.
Using the Zebrafish as a model was key to the work’s success, Boel said. As opposed to rodents, Zebrafish showed high-throughput, reproduced much faster, could be fertilized ex vivo, and could develop into maturity outside an animal uterus. “These characteristics make zebrafish the idea in vivo disease model to conduct large-scale genome editing studies,” she added.
Vetting new technologies like CRISPR in helpful animal models will be essential to determining the safety and ethical foundations of controversial uses of CRISPR, such as He Jiankui’s study on CRISPR in human embryos. He, then a scientist at Southern University of Science and Technology of China, used CRISPR to directly remove the CCR5 gene in human embryos, theoretically making the offspring (girls, in this case) resistant to HIV infection.
In an interview with the Science History Institute, CRISPR co-inventor and University of California, Berkeley, researcher Jennifer Doudna condemned He’s research as “reckless experimentation on the girls in China,” which “shattered scientific, medical and ethical norms.” In a letter to Time magazine, Doudna said, “scientists have purposely taken a cautious and deliberate approach, focusing on how to safely apply genome editing to cure genetic diseases, fight cancer, accelerate drug development, create transplant organs and develop more nutritious crops.” This cautious and deliberate approach involves using effective animal models to first determine the safety and physiological events that can happen before even considering use in humans, much less in an embryo taken to full-term. For his part, He claims he worked in mouse and monkey models before working on human embryos, but his work remains unpublished, and his first announcement to the world about his research focused, of course, on the human embryos.
- Keshavan, M. (2019). CRISPR paper that sent stocks tumbling is retracted – STAT. [online] STAT. Available at: https://www.statnews.com/2018/03/30/crispr-stocks-retraction/
- Schaefer, K., Wu, W., Colgan, D., Tsang, S., Bassuk, A. and Mahajan, V. (2019). Unexpected mutations after CRISPR–Cas9 editing in vivo. [online] Nature Methods. Available at: https://www.nature.com/articles/nmeth.4293