Off-Target Edits: How Often Does CRISPR Miss, and Does It Matter?

No molecular tool is perfectly precise, and Cas9 is no exception. The question that still keeps researchers and regulators up at night is: when it misses, how bad is the miss?

An "off-target edit" happens when Cas9 cuts a site in the genome that's similar to — but not identical to — the intended target. Human genomes are large, and sequences with 1 or 2 mismatches to the guide RNA do exist scattered throughout. Whether Cas9 cuts them depends on where the mismatches are (mismatches near the PAM end are better tolerated than mismatches at the distal end), how long Cas9 lingers on the DNA, and the concentration of the CRISPR components in the cell.

Early studies found off-target rates that worried people. Cells showed edits at multiple unintended sites. For a research tool used in cell lines, that might be acceptable. For a therapy that patients receive once and keep forever, it needs to be much better.

Several improvements have since brought off-target rates down considerably.

**High-fidelity Cas9 variants** (eSpCas9, SpCas9-HF1, HypaCas9) were engineered to reduce non-specific DNA binding. They're less likely to cut when there's a mismatch, with reduced off-target editing at most previously problematic sites while largely maintaining on-target efficiency.

**Shortened guide RNAs** (17-18 nt instead of 20) paradoxically improve specificity in some contexts — the shorter guide tolerates mismatches poorly, reducing off-target activity.

**Transient delivery methods** reduce the duration Cas9 is present in the cell. Delivering Cas9 as mRNA (rather than as a gene that keeps expressing) means editing happens once and then stops as the mRNA degrades. This is why LNP + mRNA delivery has become the clinical standard.

For Casgevy and other blood-based therapies, clinical data suggests the off-target editing profile is acceptable — regulators examined potential off-target sites computationally and experimentally and found no evidence of functional impact.

The more important concern for the future is what happens when CRISPR edits cells that then divide indefinitely. A cancer-relevant off-target edit in a stem cell could, in theory, propagate to billions of daughter cells over years. Long-term surveillance in treated patients will be essential.

Off-Target Edits: How Often Does CRISPR Miss, and Does It Matter?

// COMMENTS

ON THIS PAGE