How well does prime editing actually work in human cells?
Prime editing can be remarkably efficient in some cell types, but performance is inconsistent. In one of the most striking results, optimized all-in-one prime editing constructs achieved up to 95% editing efficiency in HEK293T cells (a standard human kidney cell line), with an average of 67% [1]. This means that in a dish, nearly every cell can be edited. However, the same study found that efficiency in K562 (blood) and HeLa (cervical cancer) cells remained low, showing that cell type matters enormously [1].
In human organoids—miniature 3D tissues that better mimic real organs—prime editing successfully modeled cancer mutations (up to 25% in colon organoids and 97% in liver organoids) and repaired the cystic fibrosis mutation CFTR-F508del [2]. Importantly, whole-genome sequencing of these edited organoids found no off-target edits, a critical safety signal [2]. Similarly, in human embryonic stem cells, prime editing achieved all types of precise edits (substitutions, insertions, deletions) without the guide-RNA-independent off-target mutations that plague some other editing tools [11].
What are the main hurdles keeping prime editing from the clinic?
Delivery is the single biggest barrier. Prime editors are large protein-RNA complexes, and getting them into the right cells in a living patient—especially hard-to-reach tissues like the retina or brain—is a major engineering challenge [6][10]. Researchers are developing virus-derived particles and adeno-associated virus (AAV) vectors to ferry the editing machinery, but these approaches still face issues with packaging capacity, immune responses, and targeting specificity [5][10]. For example, while AAV has been used to deliver prime editors into the mouse brain, translating this to humans requires overcoming significant safety and manufacturing hurdles [9].
Efficiency also drops sharply in the cell types most relevant for therapy. While editing can exceed 90% in easy-to-work-with cell lines, it is often much lower in primary cells, stem cells, and post-mitotic cells like neurons or photoreceptors [1][6][10]. A 2023 review noted that prime editing 'is still in its infancy' and that 'further development is needed to improve editing efficiency and delivery strategies for therapeutic applications' [7]. Additionally, some versions of the editor (nuclease-based prime editors) can introduce unwanted extra insertions at the target site, reducing product purity [1].
Where is prime editing showing the most promise for future treatments?
Prime editing is especially promising for diseases caused by single-point mutations or small insertions/deletions, which account for over half of all genetic disorders [3]. This includes inherited retinal diseases, cystic fibrosis, hemophilia, and certain cancers [2][4][6]. For example, prime editing has been used to functionally repair the CFTR-F508del mutation in human organoids, a direct step toward a potential therapy for cystic fibrosis [2]. In the eye, where many blinding diseases are caused by small genetic errors, prime editing has been successfully tested in retinal cells and animal models, with newer variants achieving over 50% editing efficiency in vitro [6].
The technology's safety profile is a major advantage. Because prime editing nicks only one strand of DNA instead of breaking both strands (as CRISPR-Cas9 does), it avoids the large unintended mutations and chromosomal rearrangements that can occur with double-strand breaks [3][8]. Multiple studies using whole-genome sequencing have found no off-target edits from prime editing, unlike some base editors [2][11]. This precision makes it an attractive platform for conditions where even a small error could be dangerous, such as in the brain or in stem cells [9][11].
About These Sources
This answer is built on 11 peer-reviewed studies — published from 2021 to 2026, 6 from 2024 or later, 6 in Q1 journals, collectively cited 594 times — selected as the most relevant from 11 studies that passed quality screening, drawn from 49 papers retrieved from a database of over 500 million.
Sources used in this answer
Optimized nickase- and nuclease-based prime editing in human and mouse cells
Optimized all-in-one prime editing constructs achieved up to 95% editing efficiency in HEK293T cells (mean 67%), but efficiency remained low in K562 and HeLa cells; a nuclease version improved initiation but caused extra insertions; zygotic injection in mice achieved up to 100% efficiency in fetuses.
Evaluating CRISPR-based prime editing for cancer modeling and CFTR repair in organoids
Prime editing modeled TP53 mutations in colonic organoids (up to 25% efficiency) and hepatocyte organoids (up to 97%), and repaired the CFTR-F508del mutation; whole-genome sequencing revealed no detectable off-target effects.
Prime editing: A gene precision editing tool from inception to present
Review discussing that over half of genetic diseases are caused by single-nucleotide mutations, and prime editing avoids double-strand breaks, laying a foundation for clinical application.
Gene Editing for Haemophilia-The Next Frontier.
Review noting that base and prime editors have been successfully used in cellular models to correct hemophilia gene defects, but their translational potential is limited by the diversity of causative variants.
Gene Editing by Ferrying of CRISPR/Cas Ribonucleoprotein Complexes in Enveloped Virus-Derived Particles
Review describing enveloped retrovirus-derived particles as carriers for CRISPR/Cas ribonucleoproteins, showing effective editing in primary cells and in vivo in mouse retina, liver, and brain, with potential for large-scale production.
Prime editing for ocular gene therapy and disease modeling: a narrative review of advances, delivery, and translational readiness.
Narrative review reporting that prime editing efficiency has improved from 0.7–5.5% to over 50% in vitro through optimization; highlights feasibility in retinal cells and animal models for inherited retinal diseases.
Prime editing: advances and therapeutic applications
Review summarizing that prime editing can achieve any base substitution, insertion, or deletion without double-strand breaks, but notes it is 'still in its infancy' and requires improved efficiency and delivery for therapy.
Prime editing – an update on the field
Perspective article stating that prime editing eliminates the need for double-strand breaks, increasing precision, and that advances are pushing the method toward clinical applicability.
Prime editing of the β
Study using AAV to deliver prime editors into the mouse brain to introduce a naturally occurring Adrb1 mutation, demonstrating in vivo neural editing.
Base and Prime Editing for Inherited Retinal Diseases: Delivery Platforms, Safety, Efficacy, and Translational Perspectives.
Review on base and prime editing for inherited retinal diseases, noting challenges including limited editing efficiency in photoreceptors, off-target risks, and suboptimal delivery to the outer retina.
Comprehensive analysis of prime editing outcomes in human embryonic stem cells
Comprehensive testing of prime editing in human embryonic stem cells showed successful induction of all edit types; whole-genome sequencing found no guide RNA-independent off-target mutations, unlike cytosine base editors.