NEW YORK – Researchers from Columbia University and their collaborators have developed a new genome editing tool that can integrate DNA into human cells in a targeted way without causing double-strand breaks.
Described in a Nature Biotechnology study published last month, the approach relies on a family of enzymes named CRISPR-associated transposases (CASTs), which have previously shown to be able to insert kilobase-sized DNA pieces into bacterial genomes while still maintaining programmability and accuracy.
By enabling the CAST system to work in human cells, the study paves the way for its use in research and therapeutic applications.
“I often think about [CASTs] as molecular glues,” said Samuel Sternberg, a biochemistry professor at Columbia University and the corresponding author of the study. “They are really well suited to be used for RNA-guided DNA insertion.”
Sternberg’s group first demonstrated the working mechanism of CASTs in bacterial cells in a 2019 Nature study and subsequently showed that they can be used for multiplexed, kilobase-scale genome engineering in bacteria. Since then, the group has been on a journey to repurpose CAST systems to work in human cells, he said.
Mechanistically, the CAST system employs both a CRISPR machinery, which contains a guide RNA that can recognize a matching sequence in the genome, and an integration module, which is comprised of transposases that can carry out the editing. According to Sternberg, CASTs work by introducing two nicks, or single-strand breaks, in the genome that are separated by payload, which can be “seamlessly” rejoined as the donor DNA is integrated into the genomic target site.
Unlike the conventional CRISPR-Cas9 method, the CAST approach does not lead to double-strand DNA breaks during the editing process, he said, avoiding potentially undesirable outcomes such as large-scale genomic deletions, chromosomal translocations, and chromothripsis.
The “biggest hurdle” for re-engineering CASTs for human cells was to identify specific enzymes that can function robustly in those cells, Sternberg said. To achieve that, the researchers screened about two dozen different CAST systems from a wide range of bacterial hosts, identifying a homolog from Pseudoalteromonas that was notably more active than the previous CAST systems.
Additionally, Sternberg said, the team performed a number of engineering steps to boost the CAST activity, such as fusing two of the proteins involved in CAST.
Because CAST systems are made of multiple components, the team also faced the challenge of delivering various parts of the machinery into the mammalian nucleus while ensuring they work synchronously in the right cells. To that end, the researchers identified bacterial ClpX, a sequence-specific bacterial AAA+ ATPase that enhanced genomic integration by more than two orders of magnitude.
“This is a very interesting paper,” said Magomet Aushev, a genome editing researcher at the Wellcome Centre for Mitochondrial Research at Newcastle University in the UK who was not involved in the study.
By avoiding double-strand breaks, Aushev said, the CAST system — which he referred to as “a little surgical room that does all the cutting and fixing” — can help minimize the unwanted byproducts of genome editing, potentially making the method a safer tool for therapeutic use.
In addition, Aushev applauded the technology’s ability to integrate large DNA payloads in a targeted manner into the human genome. “One of the ultimate goals of genome engineering is to be able to insert large sequences of DNA,” he said, noting that this capability can be beneficial for both therapeutic and research applications.
In their recent study, the Columbia researchers tested various CAST systems in immortalized human embryonic kidney (HEK) cells and showed that the machinery achieved single-digit percent efficiencies on plasmid substrates and low single-digit percent efficiencies on genomic target sites.
Still, Aushev said there is “a long way to go” before a CAST system can be deployed as a therapeutic tool. HEK cells are “a very established cell line, which is very easy to work with and very easy to transfect,” he pointed out, adding that CASTs might behave differently in primary patient cells or inside a living organism.
In addition, while the study has shown that CASTs are very specific, Aushev said, their efficiency is still quite low. Also, as the method’s efficiency continues to improve, its off-target effects also need to be characterized. “With genome editing, we often see that when we make the enzyme more active, not only is it active on the target you want, but it's also on [in] the other places, as well,” he explained.
“There's still a lot more work to do,” agreed Sternberg. “We have not yet fully proven that these CAST systems are safer than other methods.”
With the successful proof of concept in human cells, Sternberg said the team is now “throwing the kitchen sink at the problem” and trying to characterize CAST systems in a range of different cell types to find their best activities for downstream applications.
In addition to improving the approach’s efficiency to double-digit percentages and exploring more ways for efficient delivery, Sternberg said the team is also focusing on systematically characterizing the off-target and on-target activities of the CAST systems using sensitive, unbiased methods.
Recently, researchers at MIT developed a tool named programmable addition via site-specific targeting elements (PASTE), which combines CRISPR and integrases and also enables the targeted insertion of large DNA sequences without causing double-strand breaks.
Sternberg acknowledged that his team has “not yet carefully compared” CAST to PASTE and said it is difficult to say at this point which method is better, safer, or more efficient.
Application-wise, the team’s “biggest aspiration” is to develop the CAST technology for therapeutic use, he said. More specifically, he thinks there could be opportunities for safely and efficiently delivering an entire healthy gene into a patient’s genome for certain diseases, making CAST a universal strategy to treat the condition regardless of the type of mutation that the patient has.
Columbia University has filed a patent application pertaining to the technology, and Sternberg said the team is thinking about the best way to develop it for the benefit of future patients and researchers.
“For us, this study is a really exciting first milestone,” Sternberg said. “But it is also a starting point for work that is ongoing in the lab.”