Genetic medicine is moving from an era of cutting genes to one of rewriting incorrect DNA sequences with the desired information. Prime editing has produced meaningful results in its first patients, advancing from a laboratory technology to a clinical therapeutic platform. Now that the feasibility of precision editing has been demonstrated, the focus of competition is expanding from editing accuracy to delivery efficiency, long-term safety, and manufacturing competitiveness.

***

Gene-editing technology had long developed around the question of how precisely DNA could be cut. CRISPR gene-editing systems locate a target DNA sequence and cut both strands. As the cell repairs the damaged region, a gene can be removed or a new genetic change can be introduced. The technology was faster, less expensive, and more broadly applicable than previous methods, but the act of cutting DNA could itself create new risks. Unintended insertions or deletions could occur during the repair process, and the structure of chromosomes could potentially be altered.

Base editing reduced some of these limitations. By enabling scientists to convert a particular DNA base into another without cutting both strands, it improved the precision of genome editing. However, the combinations of bases that could be converted were limited, and it remained difficult to change several bases simultaneously or insert and delete small pieces of DNA. Because disease-causing mutations take many different forms, it was difficult to treat every genetic disorder with a single editing method.

Prime editing, first introduced in 2019, emerged to fill this gap. Often compared to the “search and replace” function in a word processor, the technology locates an incorrect segment of DNA and rewrites it with the desired genetic information. A Cas9 protein with partially restricted cutting activity, a reverse transcriptase that produces a new DNA sequence, and a prime-editing guide RNA that specifies both the target location and the intended correction operate together as a single editing system.

Its operating process also differs from conventional CRISPR gene-editing systems. Rather than cutting both DNA strands simultaneously, a prime editor makes a small incision in one strand. It then uses the information encoded in the guide RNA as a template to copy the corrected sequence into the DNA. The mutation is corrected when the cell’s repair machinery incorporates the newly synthesized sequence. The key advantages are that no separate external DNA template is required and that the technology can perform not only base substitutions but also short insertions and deletions.

In theory, prime editing could target a substantial portion of known disease-causing mutations. Its broad editing capacity, however, does not automatically translate into broad therapeutic applicability. Editing efficiency differs among cell types, and the correction rate required for a therapeutic effect varies by disease. The size and complexity of the prime-editing components also make clinical application difficult. If the initial question was whether genetic information could be altered as intended, the clinical question has become whether a sufficient amount of the editing system can be delivered safely to the necessary cells.

The clinical turning point for prime editing came from the treatment of chronic granulomatous disease, a rare immune disorder. Chronic granulomatous disease is a genetic condition in which immune cells fail to produce the reactive oxygen species needed to eliminate bacteria and fungi, causing recurrent severe infections and inflammation. In some patients, a particular mutation in the NCF1 gene deprives neutrophils of their defensive function. Existing treatment options included the long-term use of antibiotics and antifungal medications or hematopoietic stem-cell transplantation from a suitable donor.

PM359 is a prime-editing-based therapy designed to correct this mutation. CD34-positive hematopoietic stem cells are collected from the patient, the NCF1 gene is corrected in the laboratory using a prime editor, and the edited cells are returned to the patient following conditioning therapy. The concept is that once the edited stem cells become established in the bone marrow, they can continuously produce immune cells with restored function.

In April 2024, the U.S. Food and Drug Administration cleared the Investigational New Drug application for PM359. It was the first authorization allowing a prime-editing therapy to enter human clinical trials. The subsequent Phase 1/2 trial was designed to evaluate safety, biological activity, and preliminary therapeutic efficacy. A stepwise approach was adopted in which adults with stable disease were treated first, followed by patients with active infection or severe inflammation and then adolescent and pediatric patients.

Results from the first patient, released in 2025, showed that prime editing could restore clinically meaningful cellular function in humans. Before treatment, normal NADPH oxidase activity was almost entirely absent. Following treatment, however, the function was restored in 58 percent of neutrophils by Day 15 and 66 percent by Day 30. These figures substantially exceeded the threshold believed to indicate potential clinical benefit. Neutrophil and platelet engraftment also occurred relatively quickly, and no serious adverse events directly related to PM359 were reported during the initial observation period.

At the end of 2025, clinical results from two patients were published in a medical journal. The edited hematopoietic stem cells became established in the patients’ bodies, produced functional immune cells, and reportedly restored the biological activity required to defend against infection. This provided the first clinical evidence that prime editing could move beyond laboratory cells and animal models to correct the mechanism of disease in actual patients.

These results should not, however, be interpreted immediately as proof of widespread therapeutic success. The number of treated patients remained extremely small, and the observation period was limited. Researchers still need to determine whether the benefits persist for years, whether the edited hematopoietic stem cells continue producing normal blood cells over the long term, and whether unforeseen abnormalities emerge with time. The most important significance of the initial results was not the arrival of a fully established treatment but the first demonstration that prime editing could work in a clinical setting.

Prime editing was not designed as a technology limited to a single mutation or disease. By redesigning the guide RNA with a new target location and correction sequence, the platform can address various types of mutations. Rather than developing an entirely different therapeutic technology from the beginning for every disease, developers can retain the core editing system while changing the target and delivery mechanism.

There were practical reasons for beginning clinical development with blood and immune disorders. Hematopoietic stem cells can be removed from the body, edited, tested for quality, and then returned to the patient. Editing rates, cell viability, and unintended changes can be assessed before administration, allowing a relatively high degree of control. Hospitals also already possess infrastructure for hematopoietic stem-cell transplantation and cell and gene therapies. The first clinical application of prime editing therefore represented a strategic choice to demonstrate the technology’s effectiveness in cells that were comparatively manageable.

Interest in follow-up development is moving rapidly toward the liver. The liver is readily accessible through the bloodstream, and lipid nanoparticles accumulate there relatively efficiently. Genetic liver and metabolic disorders such as alpha-1 antitrypsin deficiency, Wilson disease, and phenylketonuria could produce clear biological responses when particular genetic mutations are corrected. Some diseases may also be suitable as early targets for in vivo editing because correcting only a certain proportion of liver cells could potentially produce a therapeutic effect.

A preclinical study published in 2026 used lipid nanoparticles to deliver prime-editing components to the livers of mice. The researchers corrected the mutation responsible for phenylketonuria and reduced blood phenylalanine to therapeutically relevant levels. The findings demonstrated the potential for prime editing to operate directly within an organ without removing cells from the body. Additional toxicology studies, large-animal research, and manufacturing validation will be necessary before the approach can qualify as a clinical therapy, but the technological foundation for expanding beyond ex vivo editing into in vivo treatment has become stronger.

Muscular, neurological, and retinal disorders are also considered areas for longer-term expansion. These diseases involve numerous mutations that cannot easily be corrected through a single base substitution. Further advances in the insertion and deletion capabilities of prime editing, along with twin prime editing, in which two editors work together, could make it possible to remove repeat sequences, replace exons, and modify relatively large genetic regions.

The expansion of the potential disease list, however, does not automatically mean that the clinical pipeline is expanding at the same pace. The ex vivo quality-control processes used for blood cells cannot be applied in the same way to the brain, muscles, or lungs. Each disease requires separate confirmation of how many cells must be edited to improve symptoms, whether repeat dosing is possible, and whether efficacy and safety persist as pediatric patients grow. Prime editing may be a general-purpose tool, but each therapy must still be validated separately for its particular disease and target tissue.

The greatest obstacle to the commercialization of prime editing is not its ability to rewrite genetic information but its ability to transport the editing system to the necessary cells. A prime editor is a large molecule composed of a Cas9 protein fused to a reverse transcriptase, while the guide RNA contains both the targeting instructions and the sequence to be written. Because these components are large and complex, it is difficult to package them stably within a single delivery vehicle. Even after entering a cell, they must reach the nucleus and remain active for a sufficient period.

Ex vivo editing allows many of these problems to be controlled. Editing materials can be delivered to collected cells through methods such as electroporation, after which editing efficiency and cellular condition can be evaluated so that only appropriate cells are administered to the patient. In vivo editing, by contrast, must avoid degradation in the bloodstream and reach the target tissue rather than unintended organs. After crossing the cell membrane, the delivery vehicle must release the editing materials into the cytoplasm and nucleus, and the materials should disappear quickly once their work is complete.

Lipid nanoparticles are among the most closely watched delivery systems. Manufacturing expertise has accumulated through the development of RNA medicines and vaccines, and the editing machinery can be delivered as temporarily active RNA rather than DNA. If the editing materials do not remain in the cell for an extended period, the risks of unnecessary editing and immune reactions could be reduced. Current lipid nanoparticles, however, travel efficiently to the liver but do not yet possess sufficient ability to target other tissues selectively, including the lungs, muscles, brain, and bone marrow.

Research published in 2026 jointly optimized the ratio of messenger RNA to guide RNA, the way these molecules were packaged inside particles, RNA stability, and the structure of the reverse transcriptase. New structures protecting the ends of guide RNAs reduced premature RNA degradation inside cells, while reverse transcriptases redesigned with artificial intelligence improved stability and editing activity. The field began moving beyond the isolated improvement of individual components toward designing the editor and delivery vehicle as a unified therapeutic system.

Virus-like particles, modified viral vectors, and polymeric nanoparticles are also being studied as alternatives. Virus-like particles can deliver proteins and RNA into cells while reducing the risks associated with replicating viral genetic material. Viral vectors are highly capable of reaching particular tissues, but the large size of prime editors and the possibility of prolonged expression remain concerns. Researchers are also exploring methods that divide the editor into several components for delivery and reassemble them inside the cell. However, the more complex the system becomes, the more difficult manufacturing and quality control become.

Future competition is unlikely to be judged by editing rates alone. Developers must also demonstrate whether a sufficient degree of correction can be achieved at a low dose, whether exposure in non-target tissues can be minimized, and whether particle quality can remain consistent during large-scale production. Delivery technology is becoming not merely packaging added to prime editing but a core asset that determines which diseases can be treated and how many patients can receive treatment.

Prime editing may be safer than conventional CRISPR gene-editing systems because it avoids double-strand DNA breaks. Several conditions must align before editing is completed, including the binding of the guide RNA to the target DNA, verification of the incision site, and copying of the new sequence. These multiple stages may support high specificity. The phrase “without cutting,” however, does not mean “without risk.”

Unintended results can first occur at the target site itself. The new sequence may be copied only partially, unnecessary insertions or deletions may occur, or DNA structures generated during the editing process may interact with the cell’s repair machinery in unexpected ways. The possibility of off-target editing at other sequences resembling the intended target also does not disappear completely. Even a low-frequency error could become clinically meaningful when millions or hundreds of millions of cells are treated.

The type and condition of the cells also affect safety. During laboratory culture, particular cells with strong proliferative capacity may be selected, or only cells capable of withstanding editing and conditioning procedures may survive. In hematopoietic stem-cell therapies, patients must be monitored over long periods to ensure that a specific group of edited cells does not proliferate abnormally. In vivo editing also requires confirmation that tissues other than the target organ, including reproductive cells, have not been exposed.

Immune reactions represent another important variable. Prime editors contain bacterial Cas proteins and reverse transcriptases. If the human body recognizes them as foreign materials, inflammation or other immune responses could occur, and preexisting immune memory might cause edited cells to be eliminated. If antibodies against the delivery vehicle are produced, administering the same treatment again may become difficult. This is one reason why treatment designs capable of producing a sufficient effect with a single dose are emphasized.

Safety assessment cannot be limited to examining a few predicted off-target sites. It must examine genome-wide sequence changes, large deletions and chromosomal rearrangements, cell-by-cell editing outcomes, changes in gene expression, and the possibility of tumor development. As analytical technology becomes more sophisticated, researchers may discover low-frequency by-products that previously went undetected. The discovery of more errors does not necessarily mean that the technology has regressed; it may instead indicate that the resolution of safety assessment has improved.

The focus of clinical regulation is also likely to shift from accuracy at the moment of editing toward the patient’s long-term outcome. Even after the therapeutic materials disappear from the body, the changes made to DNA remain. The absence of abnormalities during a short-term clinical trial does not eliminate the need to monitor hematological changes, tumor development, and organ function years later. Confidence in prime editing will be built not through claims that errors never occur but through transparent explanations of which risks are measured, how they are assessed, and how patients are followed over time.

Prime editing has the potential to change the development model of the pharmaceutical industry. Conventional small-molecule medicines suppress or supplement the functions of proteins associated with disease, whereas prime editing corrects the genetic information responsible for the disease itself. If a single treatment can provide a long-lasting effect, the therapeutic structure of genetic disorders that previously required lifelong medication could also change.

Industrial competition is likely to unfold across three levels. The first is competition over the editing engine. Smaller and more efficient Cas proteins, stable reverse transcriptases, and degradation-resistant guide RNAs are required. The second is competition over delivery technology. Delivery vehicles capable of selectively reaching the lungs, muscles, central nervous system, and bone marrow, rather than only the liver, must be secured. The third is competition over manufacturing and analysis. Achieving a high editing rate in a laboratory is different from repeatedly producing a uniform therapy suitable for administration to patients.

Prime editing also forces a choice between patient-specific treatment and standardized mass therapy. Even among patients with the same disease, different mutation locations may require different guide RNAs. If every mutation is developed as a separate therapy, the regulatory and manufacturing costs may become unmanageable. By contrast, developers could improve commercial viability and development efficiency by targeting mutations shared by many patients, selecting genetic regions where a single edit could address several mutations, or designing common strategies that restore gene function.

Regulators must also establish new evaluation frameworks. A central issue is whether conventional review procedures for individual medicines should be applied without modification when the editor and delivery vehicle remain the same and only the guide RNA changes. If data on the platform’s common components can be reused and only the target specificity and toxicity of the modified guide RNA require additional evaluation, therapies for rare mutations could be developed more rapidly. If the standards are too permissive, however, low-frequency genomic risks could be overlooked.

Price and accessibility are also essential conditions for clinical realization. Ex vivo treatment, in which a patient’s cells are collected, individually edited, quality-tested, and returned, involves a lengthy and expensive manufacturing process. When bone marrow conditioning and hospitalization are required, access becomes difficult in regions with limited medical infrastructure. An in vivo therapy that can be manufactured in advance could simplify costs and supply chains, but it would introduce the additional challenge of being less amenable to safety checks before administration.

The next several years will determine whether the promise of prime editing can become established as a practical therapeutic platform. Major milestones will include whether the functional recovery observed in the first patients persists over the long term, whether in vivo delivery studies progress into human clinical trials, and whether sufficient editing rates can be achieved in tissues other than the liver. Lower manufacturing costs, standardized regulation, and the accumulation of long-term safety data will also determine the technology’s pace of development.

The first clinical results from prime editing demonstrated that precisely rewriting DNA was no longer confined to theory. Genuine clinical realization, however, will not be completed by recording high editing rates in one or two patients. The field must deliver treatments of consistent quality to diverse patients, transport them safely to the necessary tissues, and demonstrate therapeutic durability and genomic stability years later. The next stage of gene therapy will be shaped not by a competition to build sharper molecular scissors, but by a competition to create systems that reach their targets more accurately, operate only as much as necessary, and track risks for as long as required.

Nature, October 2019, Anzalone, A. V., et al., Search-and-Replace Genome Editing without Double-Strand Breaks or Donor DNA

New England Journal of Medicine, December 2025, Gori, J. L., et al., Prime Editing for p47phox-Deficient Chronic Granulomatous Disease

Nature Reviews Bioengineering, September 2025, Flugel, C. L., Cadinanos-Garai, A., and Abou-el-Enein, M., A Clinical Roadmap for Base and Prime Editing

Nature Biotechnology, May 2026, Sakai, H. A., et al., Directed Evolution of Small RNA-Stabilizing Motifs That Improve Prime-Editing Efficiency

Nature Nanotechnology, June 2026, Jiang, A. Y., et al., Efficient Prime Editing In Vivo and In Vitro Using Lipid Nanoparticles