Skip to main content
Premium Trial:

Request an Annual Quote

City of Hope Team Demonstrates Strategy to Measure CRISPR, TALEN Off-target Effects

NEW YORK (GenomeWeb) – Despite their efficacy as gene-editing technologies, CRISPR/Cas9 and TALENs are still hampered by the potential for off-target activity, which can result in genome instability and impair normal gene function. But a new strategy reported this week, which had previously been developed for a related technology, promises to help researchers measure such unintended effects.

CRISPR — short for clustered, regularly interspaced, short palindromic repeats — involves the use of a nuclease known as Cas9 to induce double-strand DNA breaks (DSBs). These breaks are targeted to specific locations in the genome using a synthetic RNA that guides Cas9.

Transcription activator-like effector nucleases, or TALENs, also create targeted DSBs, but involve the use of a non-specific DNA cleavage domain fused with a DNA-binding domain.

Both technologies are widely used in research and are being explored as therapeutic modalities, but the potential for off-target effects can be an issue. As such, there remains the need for effective methods to confirm whether DNA cleavage is occurring in the right places.

Computer algorithms and in vitro selection has been used to screen for potential off-target cleavage sites, but these frequently fail to predict off-target effects observed in vivo, limiting their utility.

Aiming to develop a better method to detect off-target CRISPR/Cas9 and TALEN cleavage, City of Hope researcher Jiing-Kuan Yee and colleagues looked to a technique based on integrase-defective lentiviral vectors (IDLVs) that was originally developed by a team at the German Cancer Research Center to identify unintended cleavage sites of zinc finger nucleases (ZFNs).

The linear double-stranded IDLV genome is capable of being preferentially incorporated into DSBs during a DNA repair process known as non-homologous end-joining (NHEJ). As reported in a 2011 paper, these IDLVs can be introduced into ZFN-treated cells, essentially tagging the location of DSBs. The locations of the IDLV integration sites can then be mapped using linear amplification-mediated PCR, revealing whether the ZFNs were hitting their intended targets.

Seeing the potential in this approach, researchers led by the City of Hope's Jiing-Kuan Yee adapted it for CRISPR/Cas9 and TALENs, describing their efforts this week in Nature Biotechnology.

"It's essentially the same approach, but [we] used a different way to generate DNA double-strand breaks," Yee told GenomeWeb.

In their experiments, Yee and his colleagues generated CRISPR/Cas9 nucleases and TALENs targeting the same genomic regions in two genes known to cause disease when mutated: the Wiskott-Aldrich Syndrome (WAS) gene and the tyrosine aminotransferase (TAT) gene.

HEK cells were transfected with the expression plasmids for either the CRISPR-Cas9 nucleases or the TALENs, followed by transduction with an IDLV carrying a gene that confers resistance to the antibiotic and selective agent puromycin.

The DSBs caused by the nucleases allowed IDLV integration, resulting in a two- to threefold increase in the number of puromycin-resistant colonies, which were pooled and mapped for virus-host genome junctions.

Clustered IDLV integration sites (CLISs) within 60 base pairs of the targeted cleavage site were identified in CRISPR/Cas9- and TALEN-treated cells, but not in any cells treated with IDLVs alone. When they looked for CLISs outside of the WAS and TAT genes, the scientists found no off-target sites for the TALENs tested, but identified ones for some of the CRISPR/Cas9 nucleases.

In silico prediction and deep sequencing was used to determine the sensitivity of the IDLV assay, demonstrating that it could detect off-target cleavage with a frequency as low as one percent. This finding was validated using IDLV to test a CRISPR/Cas9 nuclease against VEGFA with known off-target sites.

Yee noted that the IDLV method is "not perfect," pointing to the assay's inability to reliably detect low-frequency off-target sites, but highlighted its unbiased nature. "And it's a biological assay; it's not based on the prediction of what can be the off-target site," he added.