Gene therapy treats disease by changing the genetic instructions inside your cells. Instead of managing symptoms with ongoing medication, it targets the root cause: a faulty or missing gene. The approach works through three basic strategies: replacing a broken gene with a working copy, switching off a gene that’s causing harm, or inserting an entirely new gene to give cells a function they didn’t have before.
That core concept is simple, but the engineering behind it involves some of the most sophisticated medicine ever developed. Here’s how each piece fits together.
Three Ways Gene Therapy Fixes a Problem
Not every genetic disease needs the same fix. Some conditions arise because a gene is missing or produces a broken protein. In those cases, delivering a healthy copy of that gene lets cells manufacture the correct protein again. This is the approach behind treatments for spinal muscular atrophy and inherited blindness, where a single defective gene is responsible for the entire disease.
Other conditions stem from a gene that’s actively doing something harmful. Here, the goal is to silence or inactivate that gene so it stops producing its damaging protein. A third strategy skips the idea of repair altogether and introduces a completely new gene, one the body has never had, to give cells a new capability. CAR-T cell therapy for certain cancers works this way: immune cells are given a gene that lets them recognize and attack tumor cells they’d normally ignore.
Getting the Gene Inside Your Cells
The biggest engineering challenge in gene therapy isn’t designing the gene. It’s delivering it to the right cells. Genes can’t simply be injected into the bloodstream and expected to find their way. They need a vehicle, and the most common vehicles are modified viruses.
Viruses are naturally built to enter cells and deliver genetic material. Scientists strip out the parts that cause illness and replace them with a therapeutic gene, turning the virus into a delivery shell. The most widely used type, called adeno-associated virus (AAV), is small and triggers very little immune response. It works especially well for reaching nerve cells, the brain, and the spinal cord. One version, AAV9, can even cross the blood-brain barrier after a single intravenous injection and produce long-term gene expression.
The tradeoff with AAV is cargo space. It can only carry a gene roughly 4.7 kilobases long, which limits which diseases it can treat. Lentiviral vectors, built from a different virus family, can carry genes nearly three times that size (up to about 15 kilobases), making them useful for larger genes. Lentiviruses are particularly effective at entering nerve cells and integrating their payload into the cell’s own DNA, which means the therapeutic gene gets copied every time the cell divides.
Non-Viral Delivery
Viruses aren’t the only option. Lipid nanoparticles, tiny fat-based capsules, can also shuttle genetic material into cells. This is the same technology behind some mRNA vaccines. Lipid nanoparticles work well for delivering gene-editing tools like CRISPR to the liver, but redirecting them to other tissues has proven difficult. They can also trigger immune reactions and carry dose-limiting toxicity.
A newer alternative uses extracellular vesicles: tiny bubbles naturally released by cells. Because they’re biological in origin, they tend to be better tolerated by the immune system and can protect their cargo from being destroyed in the bloodstream. Some, particularly those derived from stem cells, even have natural anti-inflammatory properties. This area is still early, but extracellular vesicles may eventually allow repeated doses of gene therapy without escalating immune reactions.
In Vivo vs. Ex Vivo: Two Routes of Treatment
Gene therapy reaches your cells through one of two paths. In vivo therapy delivers the corrected gene directly into your body, typically through an injection into the bloodstream or a specific tissue. You receive the viral vector, and it finds and enters the target cells on its own. Treatments for inherited blindness and spinal muscular atrophy use this approach: one injection, and the vector does its work inside the body.
Ex vivo therapy takes a longer, more hands-on route. Doctors first remove cells from your body, usually through a blood-filtering process called leukapheresis that takes four to six hours. Those cells are shipped to a specialized lab where the new gene is introduced. Once enough modified cells have been grown, they’re sent back, and you receive them through an infusion similar to a blood transfusion. The whole process, from cell collection through hospital discharge, typically takes three to eight weeks. In some cases, you’ll need chemotherapy beforehand to make room in your bone marrow for the modified cells.
CAR-T cell therapy for blood cancers follows this ex vivo model. The lab step alone, manufacturing enough engineered immune cells, can take several weeks.
Gene Editing: Rewriting DNA Directly
Traditional gene therapy adds a new gene alongside your existing DNA. Gene editing goes further: it changes the DNA sequence itself. The most well-known tool, CRISPR-Cas9, works like molecular scissors. A guide molecule leads the Cas9 protein to a precise spot in the genome, where it cuts both strands of the DNA. The cell then repairs the break, and scientists can exploit that repair process to disable a harmful gene or insert a corrected sequence.
The limitation of this “cut both strands” approach is that the cell’s repair process is somewhat error-prone. It can introduce small insertions or deletions at the cut site, which aren’t always predictable.
Base Editing and Prime Editing
Newer techniques avoid cutting the DNA entirely. Base editors use a modified, deactivated version of the Cas9 protein that binds to the target spot without slicing through both strands. Attached to this protein is an enzyme that chemically converts one DNA letter into another. Cytosine base editors convert a C-G pair into a T-A pair. Adenine base editors convert an A-T pair into a G-C pair. No break, no messy repair: just a single-letter swap.
Prime editing expands the toolkit even further. It uses a modified Cas9 that nicks only one strand of the DNA and a specialized guide molecule that carries a template of the desired edit. A reverse transcriptase enzyme then writes the new sequence directly into the genome. Prime editing can make all twelve possible letter-to-letter changes, plus small insertions and deletions, all without a double-strand break. This precision matters because it reduces the risk of unintended changes to surrounding DNA.
Risks and Safety Concerns
Gene therapy’s greatest strength, permanently altering cells, is also what makes its risks serious. The most significant danger with therapies that integrate new DNA into the genome is called insertional mutagenesis. When a viral vector inserts its payload, it can land near genes that control cell growth. If the insertion accidentally activates one of these growth-promoting genes, the affected cell can begin multiplying uncontrollably.
This isn’t theoretical. In an early clinical trial for a severe immune deficiency called SCID-X1, several children developed leukemia after gene therapy. The viral vector had inserted near a gene called LMO2, and the vector’s own regulatory elements switched that gene on, driving cancerous cell growth. In another patient from a different trial, an insertion event caused one type of blood cell to dominate the bone marrow. Research has shown that an integrated vector can influence gene expression up to 50 to 100 kilobases away from where it lands.
Immune reactions are the other major concern. Your body may recognize the viral vector as foreign and mount an inflammatory response. With non-viral delivery methods like lipid nanoparticles, toxicity tends to be dose-dependent. Newer vector designs, editing tools that avoid double-strand breaks, and biological delivery vehicles like extracellular vesicles are all partly motivated by the goal of reducing these risks.
Approved Therapies and What They Cost
Several gene therapies have received FDA approval, each targeting a different condition. Luxturna treats an inherited form of blindness caused by mutations in a single gene. Zolgensma treats spinal muscular atrophy in young children. Hemgenix targets hemophilia B. Elevidys addresses Duchenne muscular dystrophy. Multiple CAR-T cell therapies are approved for blood cancers.
These treatments are among the most expensive in medicine. Zolgensma, a one-time intravenous infusion for infants, costs approximately €1.9 million per treatment (roughly $2.1 million). The pricing reflects the economics of developing a complex biologic for a small patient population, and economic analyses suggest the minimum viable price for a manufacturer is around €1.7 million per dose. Some countries have pushed back hard: the Netherlands negotiated an 85% discount on a competing drug for the same condition, and policymakers are increasingly exploring pay-for-performance models where the full price is paid in installments, contingent on the therapy actually working over time.
For patients, the cost calculation is unusual. A single dose that costs seven figures may replace a lifetime of treatment that would ultimately cost more. That framing drives much of the debate around pricing, reimbursement, and which healthcare systems can realistically offer these therapies to the patients who need them.