Gene therapy’s history spans roughly six decades, from the first theoretical proposals in the 1960s to the FDA’s landmark approval of a CRISPR-based treatment in 2023. That arc includes dramatic highs, devastating setbacks, and long stretches of painstaking engineering work to make the core idea safe enough for routine medicine. The story is not a straight line of progress but a series of corrections, each driven by something that went badly wrong.
When the Idea First Took Shape
The notion that you could fix a disease by inserting a working copy of a broken gene into a patient’s cells grew out of basic virology research. Scientists studying cancer-causing viruses in the 1960s noticed that these viruses could splice their own genetic material into a host cell’s DNA. That observation sparked an obvious question: could you strip a virus of its harmful genes, load it with a therapeutic gene, and use it as a delivery vehicle? The concept sounded elegant, but the practical tools were nowhere near ready.
Through the late 1970s, researchers combined early methods of moving DNA into cells with the newly developed techniques of recombinant DNA technology, which let them cut and paste specific genetic sequences with increasing precision.1PubMed. An early history of gene transfer and therapy Viral genomes became the backbone of the first reliable gene-transfer systems for mammalian cells. The idea that oncogenic viruses could be re-engineered to carry therapeutic genetic material into human cells, including blood-forming stem cells, moved from speculation to serious research program.2PubMed Central. Evolution of Gene Therapy, Historical Perspective Still, more than two decades would pass between the initial concept and the first attempt in a living patient.
The First Human Trial
That first attempt came in September 1990, when researchers at the National Institutes of Health treated a four-year-old girl named Ashanthi DeSilva. She had adenosine deaminase deficiency, a rare inherited immune disorder sometimes called ADA-SCID. Without a functioning ADA gene, her body could not produce an enzyme critical for healthy immune cells, leaving her vulnerable to life-threatening infections. The team used a retroviral vector carrying a working copy of the ADA gene to modify her own white blood cells outside the body, then infused them back.3PubMed. Persistence and expression of the adenosine deaminase gene for 12 years and immune reaction to gene transfer components: long-term results of the first clinical gene therapy trial A second patient, Cynthia Cutshall, joined the trial shortly after.4PubMed. T lymphocyte-directed gene therapy for ADA- SCID: initial trial results after 4 years
The results were cautiously encouraging. Both girls showed improvement, and years of follow-up confirmed that the corrected gene persisted in their blood cells. But interpreting the outcome was complicated by the fact that both patients continued receiving the conventional treatment, a synthetic enzyme replacement, alongside gene therapy. The trial proved safety and feasibility more than it proved a cure. Even so, it opened the door to a rush of clinical enthusiasm.
The 1990s Rush and the Cystic Fibrosis Effort
The ADA-SCID trial generated enormous excitement. Within a few years, gene therapy trials were underway for dozens of diseases. Cystic fibrosis was a prime target. The gene responsible, CFTR, had been identified in 1989, and because the disease primarily damages the lungs, researchers hoped they could deliver a corrective gene directly to the airways using viruses or fatty particles called liposomes.5PubMed Central. Gene Therapy for Cystic Fibrosis Paved the Way for the Use of Adeno-Associated Virus in Gene Therapy
The cystic fibrosis story illustrates how hard gene therapy turned out to be in practice. Despite early proof that CFTR gene transfer could work in lab dishes and initial human trials, the field has run more than 30 clinical trials involving roughly 600 patients without achieving lasting clinical benefit.6PubMed Central. Gene therapy for cystic fibrosis: Challenges and prospects The lung’s natural defenses, including thick mucus and rapid cell turnover, made it extraordinarily difficult for viral vectors to reach enough cells and stay there long enough to matter. CF gene therapy remains an active area of research today, but the decades of effort underscore a lesson the field had to learn the hard way: identifying the broken gene is the easy part.
Tragedy at Penn
The hardest lesson came in September 1999 at the University of Pennsylvania. Jesse Gelsinger, an 18-year-old with a partial deficiency in the liver enzyme ornithine transcarbamylase, volunteered for a safety trial of an adenovirus-based gene therapy. He was not critically ill; his condition was manageable with diet and medication. But four days after the vector was infused into his liver, he was dead.7PubMed. Fatal systemic inflammatory response syndrome in a ornithine transcarbamylase deficient patient following adenoviral gene transfer
His body mounted a catastrophic immune reaction to the viral vector itself. Inflammatory signals surged, his organs began to fail, and the clinical team could not reverse the cascade. The 17 patients treated before him at lower doses had not experienced anything like it, nor had the animal studies predicted it. Subsequent analysis pointed to components of the viral capsid triggering an overwhelming innate immune response.8Molecular Genetics and Metabolism. Lessons learned from the gene therapy trial for ornithine transcarbamylase deficiency The case exposed how poorly researchers understood the variation in human immune responses to these vectors, and how steep the toxicity curve could be.
Gelsinger’s death did not just end one trial. It chilled the entire field. Regulatory oversight tightened dramatically. Public confidence dropped. Funding shrank. And then a second blow landed.
Leukemia in the Bubble-Baby Trials
In a French trial for X-linked severe combined immunodeficiency (SCID-X1, sometimes called “bubble baby” disease), researchers used a retroviral vector to deliver the missing gene to bone marrow stem cells. The results were initially celebrated: nine of ten patients were effectively cured of their immune deficiency. But between 31 and 68 months after treatment, four of those nine children developed T-cell leukemia.9The Journal of Clinical Investigation. Insertional oncogenesis in 4 patients after retrovirus-mediated gene therapy of SCID-X1
The cause was insertional mutagenesis. The retroviral vector had integrated near a gene called LMO2, a known cancer-promoting gene, and its powerful viral enhancer elements accidentally switched LMO2 on. That alone was not enough to cause leukemia; the cancers also acquired additional genetic mutations unrelated to the vector. But the vector insertion was the spark.10PubMed Central. Insertional mutagenesis combined with acquired somatic mutations causes leukemogenesis following gene therapy of SCID-X1 patients The SCID-X1 leukemias, combined with Gelsinger’s death just a few years earlier, created a genuine existential crisis for gene therapy. Many observers wondered whether the approach would ever be safe enough.
Building Safer Delivery Vehicles
The field’s response was to go back to the engineering bench and redesign the vectors from the ground up. One key innovation was the self-inactivating lentiviral vector. Researchers created an HIV-derived vector with a large deletion in its control region that eliminated the virus’s own promoter activity after integration. This meant the vector could no longer accidentally switch on nearby genes in the way the older retroviral vectors had done, substantially reducing the risk of insertional mutagenesis.11PubMed Central. Self-inactivating lentivirus vector for safe and efficient in vivo gene delivery Later refinements added insulator elements, DNA sequences that act like firebreaks to prevent the vector’s regulatory machinery from influencing neighboring genes. Testing showed that these newer vectors did not activate the problematic LMO2 gene that had caused leukemia in the earlier SCID-X1 trial.12Blood. A self-inactivating lentiviral vector for SCID-X1 gene therapy that does not activate LMO2 expression in human T cells
In parallel, adeno-associated virus (AAV) emerged as a workhorse delivery vehicle. AAV is a small, naturally occurring virus that does not cause disease in humans. Engineered versions carry no viral genes at all and, critically, their DNA mostly stays as a separate loop inside the cell nucleus rather than stitching itself into the patient’s chromosomes. Random integration still happens in a small fraction of cases, but at a far lower rate than with retroviruses.13PubMed Central. Adeno-Associated Virus (AAV) as a Vector for Gene Therapy Different AAV variants naturally home to different tissues, which gave researchers the ability to target the liver, the eye, the brain, or muscle depending on which version they chose. AAV became the backbone for a new generation of gene therapies.
The First Approved Products
China moved first. In 2003, its drug regulator approved Gendicine, an adenovirus carrying the tumor-suppressor gene p53, for the treatment of head and neck cancer. It reached the commercial market in 2004, making it the world’s first approved gene therapy product.14PubMed. The First Approved Gene Therapy Product for Cancer Ad-p53 (Gendicine): 12 Years in the Clinic Western regulators were more cautious. Europe’s first approval came in 2012 with Glybera, a treatment for a rare fat-metabolism disorder that was eventually withdrawn for commercial reasons. The United States did not approve its first gene therapy until 2017, when the FDA cleared Luxturna, an AAV-based treatment for a form of inherited blindness caused by mutations in the RPE65 gene.15PubMed. Voretigene Neparvovec (Luxturna) for Biallelic RPE65-Associated Retinal Dystrophy: Systematic Review Luxturna was a milestone: a one-time injection into the eye that restored functional vision in patients who had been going blind. It showed that gene therapy could deliver durable, life-changing results.
Gene Therapy Meets Cancer Immunotherapy
Around the same time, gene therapy converged with cancer immunotherapy in the form of CAR-T cell therapy. The concept had been published decades earlier, but it took thirty years of refinement before the FDA approved the first CAR-T product in 2017.16PubMed Central. The long road to the first FDA-approved gene therapy: chimeric antigen receptor T cells targeting CD19 The approach extracts a patient’s own immune cells, genetically modifies them to recognize a specific protein on cancer cells, then infuses them back to hunt down the tumor. Unlike earlier gene therapies aimed at correcting a broken gene, CAR-T cells are engineered to do something the body’s cells were never designed to do. The FDA classified them as gene therapies because the treatment hinges on genetic modification, but the purpose is cancer killing, not gene repair. Several CAR-T products have since been approved for blood cancers, and trials are underway for solid tumors.
The Rise of Genome Editing
All the therapies described so far share a basic strategy: deliver a working gene to compensate for a broken one, without actually fixing the original mutation. Genome editing changed that premise. Instead of adding a gene alongside the damaged copy, editing tools go to the precise location of the mutation and correct it in place.
The first programmable editing tools were zinc-finger nucleases, engineered proteins that could be designed to cut DNA at a chosen spot. They worked, but designing them was expensive and labor-intensive. Next came TALENs, which were easier to engineer but still cumbersome.17G3 Genes|Genomes|Genetics. Comparing Zinc Finger Nucleases and Transcription Activator-Like Effector Nucleases for Gene Targeting in Drosophila Then in 2013, multiple research groups demonstrated that a bacterial immune system called CRISPR-Cas9 could be reprogrammed with a short piece of RNA to cut virtually any DNA sequence in human cells.18PubMed. Targeted genome engineering in human cells with the Cas9 RNA-guided endonuclease The simplicity was revolutionary. Where zinc-finger nucleases required painstaking protein engineering for each new target, CRISPR required only swapping out a short RNA guide sequence. A graduate student could design a new edit in a day.
CRISPR moved from lab discovery to approved medicine in barely a decade. In December 2023, the FDA approved Casgevy (exagamglogene autotemcel), a CRISPR-Cas9-based therapy for sickle cell disease. The treatment works by editing a patient’s own blood stem cells to boost production of fetal hemoglobin, a form of hemoglobin that compensates for the defective adult version. In clinical trials, the edit reduced the painful crises that define sickle cell disease and eliminated the need for blood transfusions.19PubMed Central. Revolutionary breakthrough: FDA approves CASGEVY, the first CRISPR/Cas9 gene therapy for sickle cell disease Casgevy became the first therapy based on CRISPR to reach the market anywhere in the world.20PubMed. Sickle Cell Disease Approvals Include First CRISPR Gene Editing Therapy
Editing Without Cutting
Standard CRISPR-Cas9 works by making a clean break in both strands of the DNA double helix and letting the cell’s own repair machinery fix it. That repair process is effective but imprecise; the cell sometimes introduces small errors at the cut site. For diseases caused by a single misplaced letter in the genetic code, researchers wanted something more surgical. Base editors, first described in 2016, can chemically convert one DNA letter to another at a specific position without ever breaking both strands. Two classes exist: one converts C-to-T (and the reverse), the other converts A-to-G. Together they cover a large share of the single-letter mutations that cause genetic disease.21PubMed Central. CRISPR-Cas9 DNA Base-Editing and Prime-Editing
Prime editing, introduced in 2019, goes further still. It can make any of the twelve possible letter swaps, as well as small insertions or deletions, all without a double-strand break. In preclinical work, both base editing and prime editing have been used to correct mutations responsible for Duchenne muscular dystrophy, restoring production of the missing protein in heart cells derived from patient stem cells.22PubMed Central. Precise correction of Duchenne muscular dystrophy exon deletion mutations by base and prime editing Neither approach has reached a commercial product yet, but both are in clinical development.
A different frontier avoids changing the DNA sequence altogether. Epigenome editing uses modified CRISPR tools that have been stripped of their cutting ability and fused instead to proteins that turn genes on or off by altering chemical tags on the DNA or its packaging proteins.23Molecular Therapy. History of Gene Therapy: From Early Trials to CRISPR In Prader-Willi syndrome, for instance, patients carry an intact but silenced copy of the genes they need. Researchers have used epigenome editing to reactivate that silenced copy in patient-derived stem cells, and in some experiments a single transient treatment produced stable, long-lasting gene reactivation.24PubMed Central. Activation of the imprinted Prader-Willi syndrome locus by CRISPR-based epigenome editing The approach is also being explored for neurological and neuropsychiatric conditions, where fine-tuning gene activity rather than switching genes fully on or off may be critical.25PubMed Central. Epigenetic Editing in Neurological and Neuropsychiatric Disorders: Pioneering Next-Gen Therapeutics for Precision Gene Control
Getting the Edit Where It Needs to Go
Having a precise editing tool is only half the challenge. You also need to deliver it to the right cells inside a living person. Viral vectors remain the most common delivery vehicle, but they have limits: they can trigger immune responses, their cargo capacity is restricted, and once someone has been exposed to a particular virus, repeat dosing is difficult. Lipid nanoparticles, the same fatty-bubble technology used in the COVID-19 mRNA vaccines, offer an alternative. They can encapsulate CRISPR components and ferry them into cells without any viral material at all.26PubMed Central. Lipid nanoparticles: The game-changer in CRISPR-Cas9 genome editing
Early work has shown that lipid nanoparticles loaded with CRISPR-Cas9 can edit genes in the mouse liver with high efficiency and without detectable toxicity. In one study targeting a blood-clotting gene, the edit improved clotting function in hemophilia mice and produced no significant off-target cuts or immune reactions against the Cas9 protein.27PubMed Central. In vivo delivery of CRISPR-Cas9 using lipid nanoparticles enables antithrombin gene editing for sustainable hemophilia A and B therapy Reaching organs beyond the liver is harder, because lipid nanoparticles tend to accumulate there after intravenous injection. Recent work has identified formulations that preferentially target the lung, editing cells in the airway with no detectable off-target editing in the liver.28PubMed. Lipid Nanoparticles for In Vivo Lung Delivery of CRISPR-Cas9 Ribonucleoproteins Allow Gene Editing of Clinical Targets If lipid nanoparticles can be reliably steered to different tissues, they could eventually replace viruses as the default delivery method for gene editing.
The He Jiankui Affair and the Germline Taboo
Every gene therapy and editing treatment approved to date modifies only the patient’s own body cells. The changes are not passed to future generations. In November 2018, Chinese biophysicist He Jiankui announced that he had crossed the one line the field had broadly agreed should not be crossed: he had used CRISPR to edit human embryos, and two of those embryos had been born as living babies.29PubMed Central. CRISPR’d babies: human germline genome editing in the ‘He Jiankui affair’
He claimed the edits would make the children resistant to HIV, but the scientific community’s response was overwhelmingly negative. The experiment had questionable scientific value, an unreasonable risk-benefit ratio, an illegitimate ethics review, and flawed informed consent. It violated China’s own regulations, which prohibit clinical use of germline genome editing, as well as the broader international consensus against heritable human editing.30PubMed Central. Experiments that led to the first gene-edited babies: the ethical failings and the urgent need for better governance He was sentenced to three years in prison by a Chinese court.
The incident accelerated calls for binding international regulation. In the years since, many countries have enacted or strengthened laws restricting clinical germline editing. But there is no unified international framework, and the resulting regulatory patchwork has created ongoing tensions between safety, ethics, and the desire not to block potentially beneficial research entirely.31Politics. Global governance of human germline genome editing: An advocacy coalition framework analysis The scientific tools to edit a human embryo exist today. The societal consensus on when, if ever, it would be acceptable to use them does not.
The Access Problem
The technical progress in gene therapy has outpaced the economic and logistical infrastructure needed to deliver it widely. Many of the newest therapies carry price tags in the millions of dollars per patient. That is partly a consequence of the manufacturing process: treatments like Casgevy or CAR-T require extracting a patient’s own cells, editing them in a specialized facility, and shipping the modified cells back. Every dose is essentially custom-made. Add the costs of chemotherapy conditioning before infusion and post-treatment monitoring, and the total expense is staggering.
Even in wealthy countries, insurance coverage and payment structures are poorly suited to one-time therapies that deliver their benefit over a lifetime but demand the entire cost up front. In lower-income countries, the challenges multiply: few facilities can manufacture cell therapies, cold-chain logistics are difficult, and regulatory expertise in gene therapy is thin. Disparities in access range from financial constraints to infrastructure limitations to differences in how national regulators approve these products.32PubMed Central. How to democratize cell and gene therapy: A global approach Proposed solutions include decentralized manufacturing closer to patients, harmonized regulatory pathways across countries, and outcome-based payment models where the cost is spread over years and tied to whether the therapy actually works. None of these have been widely adopted yet, and solving the access problem is arguably now the field’s most urgent challenge.