Gene Therapy Vectors: AAV vs Lentivirus Explained

You’re reading a gene therapy press release and it says “AAV9 vector” or “lentiviral ex vivo.” You know the trial result matters, but you don’t know what the vector tells you about the company’s approach, its risks, or whether this therapy is built to scale. Every gene therapy press release mentions the vector. Most investors skip past it. That’s a mistake, because the vector determines the diseases a therapy can target, how much it costs to manufacture, and the safety risks that could sink the program.

The vector is the drug delivery system. AAV (adeno-associated virus) is for in vivo gene therapy: you inject the vector into the patient and it delivers the gene directly. Lentivirus is primarily for ex vivo gene therapy: you extract the patient’s cells, modify them in a lab with the lentiviral vector, then reinfuse them. That one distinction tells you the manufacturing model, the cost structure, and the safety profile.

The Quick Comparison

Feature	AAV	Lentivirus
Delivery	In vivo (inject into patient)	Ex vivo (modify cells in lab, reinfuse)
Genome integration	Mostly episomal (stays outside genome)	Integrates into genome (permanent expression)
Target tissues	Depends on serotype (AAV9: brain/heart; AAV8: liver; AAV5: retina)	Blood cells, stem cells (ex vivo only)
Manufacturing	Hard to scale, low yields, capacity-constrained	Established CAR-T infrastructure, but per-patient batches
Key safety risk	Immunogenicity, hepatotoxicity at high doses	Insertional mutagenesis (cancer risk)
Real example	Sarepta Elevidys (AAVrh74, DMD)	Vertex/CRISPR Casgevy (lentiviral, sickle cell)

Step 1: Find the vector type in the press release

Look for “AAV” or “adeno-associated virus” versus “lentiviral” or “retroviral.” The press release or ClinicalTrials.gov entry will state it. If it says “AAV9” or “AAVrh74,” it’s an AAV-based in vivo therapy. If it says “lentiviral vector” or “ex vivo gene-modified cell therapy,” it’s lentivirus.

Example: Ultragenyx (RARE) describes UX111 as “an in vivo AAV9 gene therapy” for Sanfilippo syndrome type A. That tells you it’s an AAV approach delivered directly to the patient, with the AAV9 serotype chosen because it crosses the blood-brain barrier to reach the brain.

Step 2: If it’s AAV, check the serotype

The serotype number tells you which tissues the vector targets, a property called tissue tropism. This is not trivia. It determines what diseases the therapy can address.

• AAV9: crosses the blood-brain barrier. Used for neurological diseases (SMA, Sanfilippo syndrome) and cardiac conditions.
• AAV8: targets the liver. Used for hemophilia and metabolic liver diseases.
• AAV5: targets the retina. Used for inherited blindness (Spark Therapeutics’ Luxturna).
• AAVrh74: targets muscle tissue. Used by Sarepta (SRPT) for Elevidys in Duchenne muscular dystrophy.

When a company announces a gene therapy program, the serotype tells you the addressable disease market before you even read the trial design.

Step 3: Determine if the approach is in vivo or ex vivo

This distinction drives manufacturing cost and scalability more than any other factor in gene therapy.

In vivo (AAV): The vector is injected into the patient. Manufacturing produces a drug product (the vector) that can theoretically be dosed to many patients from a single batch. The challenge is producing enough vector at sufficient quality.

Ex vivo (lentivirus): The patient’s cells are extracted, genetically modified in a lab using a lentiviral vector, then reinfused. Each patient gets a custom batch. This is the CAR-T model (Gilead’s Yescarta, Janssen’s Carvykti) and the gene-edited cell therapy model (Vertex/CRISPR Therapeutics’ Casgevy for sickle cell disease).

Ex vivo therapies work but are expensive. Casgevy costs over $2 million per patient. The manufacturing is per-patient, which limits scale. For more on Casgevy’s commercial trajectory, see our Casgevy 18-month post-approval analysis and FDA approval for children age 2.

Step 4: Assess manufacturing complexity

AAV manufacturing is the bottleneck of the gene therapy industry. Production yields are low, and the industry faces chronic capacity constraints. This is why companies like Sarepta carry high burn rates and why manufacturing CRLs (Complete Response Letters) are common for BLA filings. The AAV vector manufacturing market was valued at $3.2 billion in 2026 and is projected to grow at 13.4% annually through 2035, reflecting the scale-up challenge.

Lentivirus manufacturing is more established because CAR-T infrastructure already exists. But it’s limited to autologous (per-patient) batches, which caps throughput and keeps per-dose costs high.

When evaluating a gene therapy company, check whether they have in-house manufacturing or rely on contract manufacturers. Ultragenyx manufactures UX111 entirely in the U.S. at Andelyn Biosciences in Ohio and their own facility in Massachusetts. That vertical integration reduces CMC risk at PDUFA.

Step 5: Check the vector-specific safety profile

Each vector class has a distinct safety risk. Know it before you invest.

AAV: immunogenicity and hepatotoxicity. At high doses (above 5 x 10^13 vg/kg), AAV can trigger severe immune responses, liver failure, and death. The AstraGene/Audentes AT-132 trial for X-linked myotubular myopathy was halted after three children died from liver failure at high doses. In 2025, two patients died from acute liver failure after receiving Sarepta’s Elevidys at 1.33 x 10^14 vg/kg. The FDA placed Sarepta’s clinical trials on hold. These are not isolated incidents. High-dose systemic AAV is now recognized as carrying significant immunological risk, especially in patients with pre-existing liver conditions.

Lentivirus: insertional mutagenesis. Because lentivirus integrates into the genome, it can activate oncogenes and cause cancer. Bluebird bio’s Skysona (elivaldogene autotemcel) for cerebral adrenoleukodystrophy has had seven reported cases of blood cancer in treated patients. Modern lentiviral vectors include insulator elements to reduce this risk, but it has not been eliminated. This is the trade-off: integration gives you permanent gene expression, but it carries a cancer risk that episomal AAV does not.

Common Mistakes

Treating all gene therapy as the same risk profile. An AAV9 in vivo therapy and a lentiviral ex vivo cell therapy have completely different manufacturing, safety, and cost structures. Lumping them together leads to bad investment decisions.

Ignoring the serotype. If a company is using AAV8 for a neurological disease, that’s a red flag. AAV8 targets the liver, not the brain. The serotype must match the target tissue.

Underestimating AAV manufacturing risk. AAV capacity constraints have delayed multiple programs. A company with a promising Phase 2 AAV therapy but no manufacturing plan is a dilution risk. Check the 10-Q for manufacturing partnerships and capacity commitments.

Discounting insertional mutagenesis risk in lentiviral therapies. Skysona’s seven blood cancer cases show this is real, not theoretical. If a lentiviral therapy is in early trials, long-term follow-up for malignancy is required by the FDA. A clean Phase 2 does not mean the risk is gone.

Forgetting that high-dose AAV kills. Elevidys deaths occurred at doses above 10^14 vg/kg. When a company proposes systemic high-dose AAV, especially in pediatric patients with muscle diseases, the safety risk is material. The FDA has shown it will halt trials and suspend distribution.

Final Checklist

• Vector identified: AAV or lentivirus?
• If AAV: serotype noted and tissue target confirmed?
• Approach confirmed: in vivo (AAV) or ex vivo (lentivirus)?
• Manufacturing strategy checked: in-house vs. contract, capacity sufficient?
• Safety profile reviewed: hepatotoxicity (AAV) or insertional mutagenesis (lentivirus)?
• Dose level checked: is this a high-dose systemic AAV (above 5 x 10^13 vg/kg)?

For more on gene therapy investing, see our guides on CRISPR stocks and gene editing’s second wave, gene therapy catalysts coming in Q1 2027, and how to invest in biotech stocks. For Sarepta context, see our coverage of the Severino CEO transition. The FDA maintains a list of approved cellular and gene therapy products on their website.