Gene Therapy

Why gene therapy matters

• It attacks the cause, not the symptom. Most drugs manage the downstream consequences of a broken gene for a lifetime. Gene therapy aims to supply the missing instruction once. A child with spinal muscular atrophy type 1, otherwise unlikely to reach age two, can gain and keep motor milestones after a single Zolgensma infusion delivered before motor neurons are lost.
• It converted "untreatable" monogenic diseases into targets. Roughly 7,000 rare Mendelian disorders exist, most caused by a single defective gene and most with no therapy. RPE65 retinal dystrophy, hemophilia A and B, spinal muscular atrophy, adenosine deaminase SCID, and metachromatic leukodystrophy now all have approved or late-stage gene therapies where none existed a decade ago.
• CAR-T rewrote cancer immunotherapy. By ex vivo arming a patient's own T cells with a synthetic chimeric antigen receptor against CD19, therapies like Kymriah drive complete remission in a large fraction of children with relapsed B-cell acute lymphoblastic leukemia — a population that previously faced dismal odds.
• It made the genome directly editable in patients. Casgevy, approved in late 2023, uses CRISPR-Cas9 ex vivo to disrupt the BCL11A erythroid enhancer, reawakening fetal hemoglobin and functionally curing sickle cell disease — the first regulatory approval of a CRISPR medicine.
• It can be durable — sometimes for life. In non-dividing retinal cells and neurons, or in self-renewing blood stem cells that pass the correction to every daughter cell, a single treatment can last years to decades. Hemophilia B patients from early AAV liver trials have held therapeutic factor IX for over ten years.
• It reframes the economics of medicine. A one-time cure priced at $2 million challenges healthcare systems built around chronic, recurring prescriptions, forcing new outcomes-based and installment payment models — an economic problem as novel as the biology.

How gene therapy works

Every gene therapy answers three questions: what to deliver, how to package it, and where to put it. The what can be a full functional gene (gene addition/augmentation, the classic strategy for recessive loss-of-function diseases), a silencing agent such as a short hairpin RNA or antisense oligonucleotide that knocks down a toxic gain-of-function transcript, or an editing tool — CRISPR-Cas9, a base editor, or a prime editor — that rewrites the endogenous sequence in place. Gene addition does not fix the mutant gene; it simply provides a working copy that the cell transcribes and translates alongside the broken one, which is why it works cleanly for recessive disorders but not for dominant-negative ones.

The how is almost always a viral vector, because viruses spent billions of years perfecting the delivery of nucleic acid into cells. Engineers remove the virus's own replication and packaging genes — leaving only the outer capsid and the minimal cis-acting sequences — and replace the interior with the therapeutic cassette (a promoter, the transgene, and a polyadenylation signal). Recombinant AAV is a tiny non-enveloped parvovirus that carries roughly 4.7 kilobases and persists mainly as a stable, non-integrating episome in the nucleus of quiescent cells; its capsid serotype dictates tropism — AAV2 and AAV5 for retina, AAV8 for liver, AAV9 for central nervous system because it crosses the blood-brain barrier. Lentivirus, engineered from HIV-1 with its pathogenic genes deleted and split across separate plasmids for safety, integrates up to about 8 kb permanently into the host genome, making it the vector of choice when the modified cell divides. Non-viral options — lipid nanoparticles, electroporation of naked DNA or mRNA — avoid anti-viral immunity but are less efficient at reaching most solid tissues.

The where splits gene therapy into two logistics. In ex vivo therapy, cells are removed from the patient — hematopoietic stem cells or T cells — modified in a controlled culture with lentivirus, expanded, tested for integration site and copy number, and reinfused, often after conditioning chemotherapy clears space in the marrow. In in vivo therapy, the vector is administered directly — subretinally, intravenously, intrathecally, or intramuscularly — and must navigate the bloodstream and immune system to reach the right cells on its own. Once inside, the vector is endocytosed, escapes the endosome, and traffics to the nucleus, where the transgene is transcribed by host RNA polymerase II and the therapeutic protein is finally produced.

Gene addition vs gene silencing vs genome editing

Feature	Gene addition	Gene silencing	Genome editing (CRISPR)
Goal	Supply a working gene copy	Knock down a harmful transcript	Rewrite the endogenous sequence
Fixes the mutation?	No — adds a spare	No — reduces expression	Yes — corrects in place
Best for	Recessive loss-of-function	Dominant / toxic gain-of-function	Precise correction, knockout
Typical tool	AAV or lentivirus + cDNA	shRNA, siRNA, ASO, miRNA	Cas9, base editor, prime editor
Example	Luxturna (RPE65), Zolgensma (SMN1)	Patisiran (TTR amyloidosis)	Casgevy (BCL11A enhancer)
Main risk	Overexpression, cargo-size limit	Incomplete / transient knockdown	Off-target edits, large deletions

AAV vs lentivirus vs non-viral delivery

Property	Adeno-associated virus (AAV)	Lentivirus	Non-viral (LNP / naked)
Genome fate	Mostly episomal, non-integrating	Integrates into host genome	Transient (episomal / mRNA)
Cargo capacity	~4.7 kb	~8 kb	Large, flexible
Dividing cells	Diluted out over divisions	Retained in all progeny	Diluted / degraded
Integration risk	Very low	Insertional mutagenesis (SIN-mitigated)	Minimal
Pre-existing immunity	High (30–70% of adults)	Low	Low
Typical use	In vivo — retina, liver, CNS, muscle	Ex vivo — HSCs, CAR-T	mRNA vaccines, liver siRNA
Approved example	Zolgensma (AAV9), Luxturna (AAV2)	Kymriah, Zynteglo	Patisiran (LNP)

Common misconceptions

• Gene therapy edits your DNA. Most approved gene therapies do not edit anything. Gene-addition products like Luxturna and Zolgensma leave the mutant gene untouched and simply add a working copy that sits in the nucleus as a separate episome. Only editing-based therapies (Casgevy) actually change the endogenous sequence — and even then only in the targeted somatic cells.
• It changes your children. Every approved gene therapy is somatic — it modifies retina, liver, muscle, or blood cells and is not passed on. Altering the germline (sperm, eggs, embryos) would be heritable and is prohibited in human clinical practice worldwide.
• You can just re-dose if it wears off. Re-dosing AAV is usually impossible: the first exposure raises neutralizing antibodies against the capsid that inactivate any second dose of the same serotype. This is why durability and getting it right the first time matter so much.
• Viral vectors give you the disease the virus causes. The vectors are gutted of their pathogenic genes and cannot replicate. AAV is derived from a virus with no known human disease; lentiviral vectors are built from HIV-1 but with the virulence genes removed and the components split across plasmids so no functional virus can reassemble.
• Gene therapy is brand new. The concept was articulated in the 1970s, the first authorized human trial (for ADA-SCID) ran in 1990, and the field was nearly killed by the 1999 death of Jesse Gelsinger and by leukemias in a French SCID-X1 trial. Today's successes rest on three decades of hard lessons about immunity and vector safety.
• CAR-T is a drug you take. CAR-T is a living gene therapy: a patient's own T cells are extracted, transduced with a lentivirus encoding a chimeric antigen receptor, expanded, and reinfused as a self-replicating cellular product that hunts CD19-positive tumor cells.

History and famous milestones

• The first authorized human gene therapy (1990). W. French Anderson, Michael Blaese, and Kenneth Culver treated a four-year-old, Ashanti DeSilva, for adenosine deaminase deficiency SCID by inserting a functional ADA gene into her T cells with a retroviral vector ex vivo. It proved the principle was survivable and, combined with ongoing enzyme therapy, clinically meaningful.
• The Jesse Gelsinger tragedy (1999). An 18-year-old with a mild partial ornithine transcarbamylase deficiency died four days after a high-dose adenoviral vector infusion triggered a massive systemic inflammatory response and multi-organ failure. His death halted many trials and forced a hard reckoning with vector immunogenicity and informed consent that reshaped the entire field.
• SCID-X1 and insertional oncogenesis (early 2000s). Alain Fischer's Paris team cured children of X-linked SCID with a gammaretroviral vector, but several later developed T-cell leukemia because the vector integrated near the LMO2 proto-oncogene. The disaster drove the shift to self-inactivating (SIN) lentiviral vectors with safer integration profiles.
• Luxturna and the first US in vivo approval (2017). Voretigene neparvovec, a subretinal AAV2 carrying RPE65, restored functional vision in patients with a blinding retinal dystrophy — the first directly-administered gene therapy approved by the FDA, validating the AAV platform for inherited disease.
• Zolgensma and the price of a cure (2019). A single intravenous AAV9 dose delivering SMN1 transformed spinal muscular atrophy type 1 from a lethal infantile disease into a treatable one — and, at roughly $2.1 million, became the emblem of the one-time-cure pricing debate.
• Casgevy, the first CRISPR medicine (2023). Developed by Vertex and CRISPR Therapeutics from the science of Jennifer Doudna and Emmanuelle Charpentier's 2012 CRISPR-Cas9 work, exagamglogene autotemcel ex vivo edits the BCL11A enhancer in a patient's own stem cells to switch fetal hemoglobin back on, functionally curing sickle cell disease and transfusion-dependent beta-thalassemia.