Genetic engineering encompasses a growing set of tools that allow scientists to read, cut, rewrite, or silence stretches of DNA in living organisms, and its medical applications have moved from theoretical promise to approved treatments within the last decade. The toolkit now ranges from older methods like restriction enzymes and viral gene delivery to newer precision systems like CRISPR-Cas9, base editors, and prime editors, each suited to different therapeutic problems. What makes the current moment distinctive is not just the sophistication of the editing itself but the pace at which these techniques are reaching patients with previously untreatable diseases.
Cutting DNA on Purpose
The ability to cut DNA at specific locations is the foundation of nearly every genetic engineering technique. Restriction enzymes, proteins that bacteria naturally use to chop up invading viral DNA, were the original workhorses. First demonstrated as a practical tool in 1971, these enzymes enabled scientists to slice DNA at predictable sequence motifs and rearrange the fragments, giving rise to recombinant DNA technology and, eventually, an entire biotechnology industry.1Europe PMC / PNAS. How restriction enzymes became the workhorses of molecular biology Restriction enzymes remain essential in laboratory cloning and diagnostics, but they lack the precision needed to edit a single gene inside a living human cell.
That precision arrived with programmable nucleases, most famously CRISPR-Cas9. The system uses a short guide RNA to steer the Cas9 protein to a matching DNA sequence, where it creates a double-strand break. The cell then repairs the break through one of two main pathways. One pathway joins the broken ends back together in a sloppy fashion that often introduces small insertions or deletions, effectively knocking the gene out. The other pathway uses a template to repair the break faithfully, allowing researchers to swap in a corrected sequence.2Scientific Reports. Systematic quantification of HDR and NHEJ reveals effects of locus, nuclease, and cell type on genome-editing The balance between these two outcomes has been a persistent challenge: the sloppy pathway tends to dominate, which is fine when you want to disable a gene but problematic when you want a precise correction. Recent work on engineered Cas9 variants has shown that reshaping the structure of the DNA break can suppress the sloppy pathway and shift repair toward the precise one, improving the odds of clean edits in both dividing and non-dividing cells.3PubMed Central. Altered DNA repair pathway engagement by engineered CRISPR-Cas9 nucleases
Editing Without Breaking
Double-strand breaks are powerful, but they carry risks: unwanted insertions, deletions, or rearrangements at the cut site. Two newer approaches sidestep the break entirely. Base editing fuses a deactivated or partially deactivated Cas9 to a chemical enzyme that directly converts one DNA letter into another at a targeted spot, without ever cutting both strands of the double helix.4PubMed Central. Editing the Genome Without Double-Stranded DNA Breaks The first base editors converted C·G pairs to T·A pairs.5Nature. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage A second generation, adenine base editors, flipped A·T pairs to G·C with roughly 50% efficiency in human cells, high product purity above 99.9%, and strikingly low rates of unintended insertions or deletions.6Nature. Programmable base editing of A·T to G·C in genomic DNA without DNA cleavage Together, these two classes cover all four single-letter “transition” mutations, which account for a large share of known disease-causing variants.
Prime editing pushes further. It can install not only single-letter changes but also small insertions and deletions, all without a double-strand break and without needing a separate donor template. The system pairs a Cas9 nickase (which cuts only one DNA strand) with a reverse transcriptase enzyme. A specially designed guide RNA both directs the editor to the target and carries the sequence of the desired edit, which the reverse transcriptase writes directly into the genome.7PubMed Central. Engineered pegRNAs improve prime editing efficiency Structural studies have clarified how the Cas9 nickase and reverse transcriptase coordinate during this process, which is guiding efforts to make the system more efficient.8Nature. Structural basis for pegRNA-guided reverse transcription by a prime editor
Getting the Edit to the Right Cells
An editing tool is only as useful as your ability to deliver it where it needs to go. In medicine, delivery divides broadly into two strategies. In the ex vivo approach, cells are removed from a patient, edited in the lab, and infused back. This works well for blood cells and immune cells. In the in vivo approach, the editing machinery is delivered directly into the body, which is necessary for reaching organs like the liver, brain, or muscle.9PubMed. The Future of Gene Therapy: A Review of In Vivo and Ex Vivo Delivery Methods for Genome Editing-Based Therapies
Adeno-associated viruses, or AAVs, are the most widely used viral delivery vehicles. These small, non-disease-causing viruses can be engineered to carry therapeutic DNA into target cells. Different AAV types naturally favor different tissues: some home in on the liver, others on muscle or the central nervous system, depending on the surface proteins they use to latch onto cells.10Signal Transduction and Targeted Therapy. Adeno-associated virus as a delivery vector for gene therapy of human diseases AAV does not typically integrate its payload into the host chromosome, which reduces some risks but also means the delivered gene can be lost as cells divide.
Lipid nanoparticles, the same basic technology behind mRNA COVID-19 vaccines, offer a non-viral alternative. These tiny fat-based spheres can encapsulate mRNA encoding Cas9 along with guide RNA, delivering the editing components transiently so they do their work and then degrade. In mouse studies, lipid nanoparticles carrying CRISPR cargo have achieved efficient gene editing in the liver after a single intravenous dose.11PubMed Central. Lipid nanoparticle-mediated codelivery of Cas9 mRNA and single-guide RNA achieves liver-specific in vivo genome editing of Angptl3 Researchers have also developed targeted versions of these nanoparticles that can seek out specific cell types, including blood stem cells, using surface markers like CD45 as a homing signal. In one proof-of-concept study, a single injection into fetal mice delivered CRISPR cargo to blood stem cells and achieved lasting gene editing.12PubMed Central. In utero delivery of targeted ionizable lipid nanoparticles facilitates in vivo gene editing of hematopoietic stem cells
Treating Blood Disorders With CRISPR
The first CRISPR-based therapy to reach patients targeted sickle cell disease and transfusion-dependent beta-thalassemia, two inherited blood disorders caused by defective hemoglobin. The approach does not fix the mutated gene directly. Instead, it knocks out a genetic switch that normally keeps fetal hemoglobin turned off in adults. With that switch disabled, edited blood stem cells produce high levels of fetal hemoglobin, which compensates for the faulty adult version. In a clinical study, patients with both diseases had roughly 80% of alleles edited at the target site, with no detectable off-target editing. More than a year after treatment, both patients were transfusion-independent, and the patient with sickle cell disease had no further pain crises.13PubMed. CRISPR-Cas9 Gene Editing for Sickle Cell Disease and β-Thalassemia This therapy, later approved under the name Casgevy, marked a turning point: proof that CRISPR editing could produce durable clinical benefit.
Gene Replacement for Spinal Muscular Atrophy
Not every genetic disease needs a gene edit. Some need a gene replacement. Spinal muscular atrophy, or SMA, is caused by a missing or broken copy of the SMN1 gene, which motor neurons need to survive. Without it, infants progressively lose muscle control. Onasemnogene abeparvovec uses an AAV9 vector to deliver a working copy of SMN1 in a single intravenous infusion. Clinical trials showed that a single dose improved motor function in the majority of infants treated.14PubMed Central. Recombinant Adeno-Associated Virus Serotype 9 Gene Therapy in Spinal Muscular Atrophy The AAV9 vector’s ability to cross from the bloodstream into the central nervous system is what makes this approach possible. Preclinical work continues to refine these vectors, with second-generation designs showing better rescue of motor neuron size in mouse models, and long-term safety studies confirming sustained gene expression in the brain, heart, liver, and muscle without observable toxicity.15PubMed. Long-Term Safety and Efficacy of AAV9 Vectors Expressing Human SMN1 Gene: A Preclinical Study
Editing Genes Inside a Living Person
The most dramatic recent development is in vivo gene editing: injecting CRISPR components directly into a patient’s bloodstream and editing genes inside their body. The first disease targeted this way was transthyretin amyloidosis, a condition in which a misfolded liver protein accumulates in the heart and nerves. A treatment called NTLA-2001, which uses lipid nanoparticles to deliver Cas9 mRNA and a guide RNA targeting the TTR gene in the liver, produced rapid and deep reductions in the disease-causing protein. In an early trial at a higher dose, patients saw an average reduction of roughly 87% in serum TTR protein within 28 days of a single infusion.16PubMed. CRISPR-Cas9 In Vivo Gene Editing for Transthyretin Amyloidosis Later data from expanded dosing showed mean reductions exceeding 90%, sustained through at least four to six months.17PubMed Central. Lessons from the first-in-human in vivo CRISPR/Cas9 editing of the TTR gene by NTLA-2001 trial in patients with transthyretin amyloidosis with cardiomyopathy
Researchers have also tested an alternative CRISPR system, Cas3, which creates large targeted deletions rather than small cuts. In a mouse model of the same disease, a single lipid nanoparticle treatment achieved about 49% editing in the liver and reduced serum TTR levels by roughly 80%.18Nature Biotechnology. CRISPR–Cas3-based editing for targeted deletions in a mouse model of transthyretin amyloidosis Having multiple CRISPR systems available for in vivo work gives developers options depending on whether a small disruption or a larger deletion is therapeutically preferable.
Engineered Immune Cells for Cancer
CAR-T cell therapy takes a different angle on genetic engineering. Rather than fixing a patient’s defective gene, it engineers a patient’s immune cells to recognize and kill cancer. T cells are collected from the patient, genetically modified to express a synthetic receptor that targets a protein on the surface of cancer cells, expanded in the laboratory, and infused back. This approach has been particularly successful against certain blood cancers, with six CAR-T products approved by the FDA for hematological malignancies.19PubMed Central. Chimeric Antigen Receptor T-Cells: An Overview of Concepts, Applications, Limitations, and Proposed Solutions
The technology still faces real limitations. Severe toxicities, including dangerous immune overreactions, remain a risk. In solid tumors, CAR-T cells struggle to infiltrate the hostile microenvironment around the tumor, and cancer cells can evade the engineered receptor by shedding the targeted surface protein.20Blood Cancer Journal. CAR-T cell therapy: current limitations and potential strategies Engineering advances, including fourth-generation designs that incorporate additional signaling elements, aim to improve persistence and effectiveness.21PubMed Central. Recent advances in CAR-T cell engineering
A major practical constraint is that standard CAR-T therapy is autologous, meaning each patient’s cells must be individually harvested and engineered. This is expensive and time-consuming. Allogeneic (off-the-shelf) CAR-T cells, made from healthy donor T cells, could solve this. The catch is graft-versus-host disease, where donor T cells attack the recipient’s body. Gene editing offers a workaround: knocking out the donor T cells’ own T-cell receptor prevents them from recognizing and attacking the patient’s tissues. Clinical trials of several such products, using both TALEN and CRISPR-Cas9 editing to disrupt the receptor gene, have reported either no graft-versus-host disease or only mild skin reactions.22PubMed Central. Addressing graft-versus-host disease in allogeneic cell-based immunotherapy for cancer
Gene-Edited Pig Organs for Transplantation
The shortage of human donor organs has driven interest in xenotransplantation, transplanting organs from animals into people. Pig organs are the closest match in size and physiology, but unmodified pig tissue triggers violent immune rejection. Genetic engineering addresses this by deleting pig genes that produce the most provocative surface sugars and inserting human genes that dampen the immune response.23PubMed Central. Current Techniques of Gene Editing in Pigs for Xenotransplantation The scale of editing involved is extraordinary. In 2025, a gene-edited pig kidney with 69 genomic modifications, including deletion of three major sugar antigens, inactivation of porcine retroviruses, and insertion of seven human immune-regulatory genes, was transplanted into a living human patient with end-stage kidney disease.24Nature Medicine. Immune profiling in a living human recipient of a gene-edited pig kidney Meanwhile, a Chinese research group has produced hundreds of cloned pigs carrying ten gene edits designed for xenotransplantation, including first-generation offspring, demonstrating that large-scale breeding of heavily edited donor animals is feasible.25Protein & Cell. Specific pathogen free ten gene-edited donor pigs for xenotransplantation
Epigenetic Editing and mRNA Therapeutics
Not all genetic interventions need to change the DNA sequence. Epigenetic editors use a deactivated Cas9 fused to enzymes that add or remove chemical tags on DNA or its packaging proteins, dialing gene activity up or down without altering a single nucleotide.26PubMed Central. CRISPRoff epigenetic editing for programmable gene silencing in human cells without DNA breaks The appeal is reversibility and reduced risk: no DNA breaks, no permanent sequence changes. Early results have been mixed, however. In some studies, targeted addition of repressive chemical marks succeeded in reducing gene expression, but the degree of silencing varied considerably depending on the gene and cell type, and at certain targets the chemical marks alone were not sufficient to shut a gene down.27Nucleic Acids Research. dCas9-based epigenome editing suggests acquisition of histone methylation is not sufficient for target gene repression Making epigenetic editing robust and predictable remains an active research problem.
On a parallel track, mRNA therapeutics have emerged as a form of transient genetic medicine. Rather than editing the genome, synthetic mRNA instructs cells to produce a specific protein temporarily. The mRNA degrades within days, so the effect is inherently self-limiting. Advances in mRNA design and delivery have made it possible to produce functional proteins, antibodies, and peptides inside the body, enabling rapid vaccine development and, increasingly, therapeutic applications beyond vaccines.28PubMed Central. mRNA therapeutics: Transforming medicine through innovation in design, delivery, and disease treatment The mRNA approach also underpins the lipid-nanoparticle delivery of Cas9 used in the in vivo editing trials described earlier, where the editing protein is expressed transiently from mRNA rather than from a permanent DNA template.
Off-Target Editing and How It Gets Caught
The biggest safety concern with any DNA-cutting approach is off-target editing: the possibility that the system cuts or modifies a site other than the intended one. Because CRISPR uses a short RNA guide to find its target, partial matches elsewhere in the genome can sometimes get nicked or cut too. Considerable effort has gone into developing detection methods to find these unintended edits. Genome-wide assays now exist to catalog off-target sites, and newer tools can identify not just off-target cuts but also DNA translocations, where broken chromosome ends rejoin incorrectly. These translocation events are considered more dangerous than simple off-target mutations and were historically hard to spot.29Nature Communications. PEAC-seq adopts Prime Editor to detect CRISPR off-target and DNA translocation Software tools have also been developed to statistically quantify off-target activity and translocation frequencies from sequencing data, giving researchers a clearer picture of what happens genome-wide after editing.30Nature Communications. CRISPECTOR provides accurate estimation of genome editing translocation and off-target activity from comparative NGS data
These detection advances have in turn driven improvements to the editing tools themselves. High-fidelity Cas9 variants, more specific guide RNA designs, and the breakless approaches of base and prime editing all reduce off-target risk. Still, the field remains cautious. Each new therapy undergoes extensive profiling for unintended edits before entering clinical trials, and long-term patient follow-up is standard.31PubMed Central. Off-target effects in CRISPR/Cas9 gene editing
Manufacturing Bottlenecks and Cost
A therapy that works brilliantly in a clinical trial still has to be manufactured at scale and priced affordably, and gene therapies have struggled on both counts. AAV vectors, the most common delivery vehicle for gene replacement, face persistent manufacturing challenges including low production yields, difficulty scaling up from laboratory to commercial production, and high levels of impurities that must be removed.32PubMed. Critical challenges and advances in recombinant adeno-associated virus (rAAV) biomanufacturing These issues directly inflate the price of gene therapy drugs, some of which carry list prices above one or two million dollars per patient.33PubMed Central. Advancing AAV vector manufacturing: challenges, innovations, and future directions for gene therapy
The economics create a tension that the field has not resolved. Gene therapies often target rare diseases with small patient populations, making the per-patient cost high even if manufacturing were perfect. Health systems face short-term financial shocks from one-time treatments whose long-term durability is still uncertain.34PubMed Central. Managing the challenges of paying for gene therapy: strategies for market action and policy reform in the United States Several gene therapies have already been withdrawn from certain markets not because they failed clinically but because manufacturers could not sustain the economics. Pay-for-performance agreements, outcomes-based contracts, and installment payment models are all being explored, but no single financing model has emerged as the clear solution.35PubMed Central. Challenges for gene therapy in the financial sustainability of health systems: a scoping review
The Germline Question
Every approved gene therapy and every clinical trial discussed above involves somatic editing, meaning changes to the patient’s body cells that are not passed on to their children. Germline editing, which would alter eggs, sperm, or embryos and thereby affect all future generations descended from that individual, occupies fundamentally different ethical territory.36American Journal of Human Genetics. Human Germline Genome Editing: An ASHG Position Statement For more than fifty years, the line between somatic and germline modification served as the central ethical boundary in the debate over human gene editing. That line has grown less stable in recent years, partly because the same tools used for somatic editing could, in principle, be applied to embryos, and partly because some somatic therapies administered early in development could inadvertently reach germ cells.37PubMed Central. Setting ethical limits on human gene editing after the fall of the somatic/germline barrier
The field’s early history also informs current caution. Gene therapy suffered serious setbacks in the late 1990s and early 2000s, including a patient death in an adenoviral vector trial and cases of leukemia caused by retroviral vectors inserting near cancer-promoting genes. Those events led to sweeping improvements in safety oversight, vector design, and long-term patient monitoring that underpin today’s clinical programs.38PubMed Central. Clinical Development of Gene Therapies: The First Three Decades and Counting The memory of those failures is a useful counterweight to the field’s current optimism. It is a reminder that powerful tools demand proportionally rigorous safeguards, especially as the technology moves toward broader diseases, larger patient populations, and edits that are harder to reverse.