Why is prime editing efficiency still too low for the clinic?
The single biggest evidence gap is that prime editing simply doesn't work often enough in the cells that matter for therapy. In a 2025 study that successfully corrected a heart disease mutation in mouse embryos, the editing efficiency was just 9.3% in collected embryos and only 7.3% in the fetal hearts that actually needed fixing [2]. That means more than 90% of cells were left unedited. A 2023 review sums up the field's consensus: 'the effectiveness of the method is not yet high enough to apply it in clinical trials' [6]. A 2024 review agrees, calling efficiency a 'limitation' that needs improvement before prime editing can treat hereditary diseases [7].
Efficiency isn't just low on average—it's also highly variable depending on the cell type and the specific edit. The original 2021 study that discovered the mismatch repair problem tested 191 edits across seven mammalian cell types and found that editing outcomes varied wildly [1]. A 2021 toolkit for bacteria reported that while 1-base-pair deletions worked 40% of the time, larger edits (like inserting 97 base pairs) 'dropped sharply' in efficiency [3]. This variability makes it hard to predict whether prime editing will work for a given patient's mutation.
How does the cell's own repair system sabotage prime editing?
A major hidden obstacle is the cell's DNA mismatch repair (MMR) system, which actively fights prime editing. In a landmark 2021 study, researchers used genetic screens to discover that MMR proteins recognize the mismatched DNA created during prime editing and try to 'fix' it—often by deleting the edit or creating unwanted insertions or deletions (indels) [1]. By temporarily blocking MMR with an engineered protein, they boosted prime editing efficiency by an average of 7.7-fold compared to the standard system, and improved the ratio of correct edits to unwanted byproducts by 3.4-fold [1]. This finding directly explains why efficiency is so cell-type dependent: cells with high MMR activity (like stem cells) are naturally harder to edit.
The same study also found a clever workaround: strategically placing silent mutations (changes that don't alter the protein) near the intended edit can help the edit evade MMR detection [1]. This 'evasion' strategy is now being used in practice—the 2025 heart disease study deliberately introduced a silent mutation alongside the therapeutic correction to distinguish edited cells from wild-type ones [2]. So the evidence shows that MMR is not just a minor nuisance; it's a fundamental biological barrier that must be managed for prime editing to work reliably.
Why are delivery and guide RNA stability such stubborn problems?
Prime editing requires delivering a very large protein (Cas9 nickase fused to reverse transcriptase) plus a specially designed guide RNA (pegRNA) into the right cells. Multiple reviews highlight delivery as a 'big challenge' because the prime editor components are too large for standard viral vectors like AAV [6][7]. A 2025 study that successfully delivered prime editors into the mouse brain using AAV vectors is a notable exception, but it remains a proof-of-concept rather than a routine method [8]. The 2024 review bluntly states that 'delivery of prime editors is also a big challenge due to their size' [7].
Even when delivery succeeds, the guide RNA itself is fragile. A 2021 study showed that the 3' end of the pegRNA—the part that carries the reverse transcriptase template and primer binding site—is vulnerable to degradation, which 'poisons' editing activity [5]. By adding structured RNA motifs to the 3' end to stabilize it, the researchers boosted editing efficiency 3-4-fold in multiple human cell lines without increasing off-target effects [5]. This is one of the few bottlenecks with a clear, simple fix (engineered pegRNAs, or epegRNAs), but it still hasn't been widely adopted in therapeutic studies—the 2025 heart disease study, for example, used the standard PE3 system without epegRNAs [2].
A 2022 study took a different approach, testing 12,000 different peptide fusions to find ones that boost prime editing efficiency [4]. They found that certain peptides work by increasing translation of the prime editor protein, and combining the best peptides gave further gains [4]. This suggests that multiple bottlenecks—RNA stability, protein expression, and cellular repair—can be tackled simultaneously, but no single solution has yet closed the efficiency gap enough for clinical use.
About These Sources
This answer is built on 8 peer-reviewed studies — published from 2021 to 2025, 3 from 2024 or later, 6 in Q1 journals, collectively cited 1,724 times — selected as the most relevant from 11 studies that passed quality screening, drawn from 55 papers retrieved from a database of over 500 million.
Sources used in this answer
Enhanced prime editing systems by manipulating cellular determinants of editing outcomes
Discovered that DNA mismatch repair (MMR) impedes prime editing and promotes indels; blocking MMR with an engineered protein boosted efficiency by an average of 7.7-fold across 191 edits in seven cell types.
Prime editing correction of an Nkx2-5 variant in a murine congenital heart disease model
First successful correction of a congenital heart disease mutation in mouse embryos using prime editing, but efficiency was only 9.3% in embryos and 7.3% in fetal hearts.
A versatile genetic engineering toolkit for E. coli based on CRISPR-prime editing
Developed a prime editing toolkit for E. coli achieving up to 40% efficiency for 1-bp deletions, but efficiency dropped sharply for larger edits (e.g., 97-bp deletions).
Peptide fusion improves prime editing efficiency
Screened 12,000 peptide fusions and identified peptides that enhance prime editing efficiency by increasing translation of the prime editor protein.
Engineered pegRNAs improve prime editing efficiency
Showed that degradation of the 3' end of pegRNA poisons prime editing; adding structured RNA motifs (epegRNAs) improved efficiency 3-4-fold in multiple human cell lines.
Prime editing: advances and therapeutic applications
Review concluding that prime editing efficiency and delivery strategies need further development before therapeutic applications are feasible.
Evolution of Prime Editing Systems: Move Forward to the Treatment of Hereditary Diseases
Review stating that prime editing efficiency is not yet high enough for clinical trials and that delivery is a major challenge due to the large size of prime editors.
Prime editing of the β
Demonstrated delivery of prime editors into the mouse brain using AAV vectors to introduce a naturally occurring mutation, showing feasibility but not routine efficiency.