Adeno-associated virus (AAV) transduction powers gene therapy by using a small, nonpathogenic parvovirus capsid to package a therapeutic DNA cassette, bind specific cell-surface receptors, enter through endocytosis, escape the endosome, traffic to the nucleus, and deliver single-stranded DNA that is converted to double-stranded DNA and expressed as protein. The process is a sequence of binding, uptake, escape, nuclear entry, and expression, each a bottleneck that determines dose and outcome. In practice, the capsid choice and the DNA design together set which tissues get treated.
Binding and entry: how the capsid picks its cells
AAV transduction begins with attachment. The capsid first binds heparan sulfate proteoglycans or other glycans, then engages a proteinaceous receptor. AAV serotypes differ in this tropism: AAV2 uses heparan sulfate as its primary attachment factor, while AAV9 binds galactose moieties and shows broad systemic distribution. AAV5 uses sialic acid. These differences are the foundation of tissue targeting, and they explain why "which AAV" is the first manufacturing decision.
After attachment, the virus is internalized by clathrin-mediated endocytosis, though alternative routes exist. Inside the endosome, the low pH triggers a conformational change in the capsid. This exposes a phospholipase domain in the VP1 protein, which remodels the endosomal membrane and allows the capsid to escape into the cytoplasm. Without this escape, the DNA never reaches the nucleus.
The nuclear step is rate-limiting. The intact capsid must enter through the nuclear pore. Whether AAV relies on nuclear pore import or waits for nuclear envelope breakdown during mitosis depends on the target tissue. In post-mitotic cells such as neurons and hepatocytes, intact capsid entry through the pore is essential. In dividing cells, breakdown of the nuclear envelope during mitosis can provide an alternate route.
From capsid to protein: the intracellular sequence
Inside the nucleus, the single-stranded DNA genome must be converted to double-stranded DNA before transcription. AAV is a single-stranded DNA virus, and the cell's own machinery performs second-strand synthesis. This conversion is slow and is often the rate-limiting step in expression. Co-infection with a helper virus or DNA-damage stress can accelerate it.
The therapeutic cassette contains a promoter, the transgene, and regulatory elements such as polyadenylation signals. AAV genomes are small. The packaging limit is roughly 4.7 kilobases of single-stranded DNA, including the two inverted terminal repeats (ITRs) that flank the cassette. Larger inserts produce truncated or fragmented genomes that package inefficiently. This is the core engineering constraint that shapes every vector design.
The ITRs are the only viral sequences retained. The rep and cap genes are supplied in trans during manufacturing, so the vector cannot replicate on its own. This is a deliberate safety design and a reason AAV is considered nonpathogenic.
After expression, episomal DNA persists. AAV delivered DNA generally remains as an episome, a circular concatemer that does not integrate into the genome at high frequency. This supports durable expression in non-dividing cells but means expression can dilute in actively dividing tissue. Rare integration events do occur, and they are a long-term safety question rather than a certainty of either benefit or harm.
A simple way to picture it: the capsid is a delivery truck, the ITRs are the shipping labels, and the transgene is the cargo that stays in the cell's nucleus.
Quantified context: dosing and kinetics
AAV doses in clinical use are not trivial. Systemic intravenous administration typically ranges from about 1 to 2 times 10^13 vector genomes per kilogram of body weight. Direct local administration to the eye or central nervous system uses lower absolute doses but still in the range of 10^11 to 10^12 vector genomes. These are order-of-magnitude figures and vary by product and indication.
The body has a finite pool of receptors and a clearance system. A significant fraction of systemically administered capsid is cleared by the liver and spleen, and pre-existing neutralizing antibodies can bind and inactivate the vector before it reaches target cells. This is why many gene therapy protocols screen patients for pre-existing anti-AAV antibodies.
Expression kinetics vary with route. Intravitreal injection into the eye can produce detectable transgene expression within weeks. Intravenous liver-directed therapy may take weeks to months to reach steady-state protein levels. Systemic administration also activates the complement and innate immune systems, and high doses are associated with the risk of liver enzyme elevation and, rarely, more serious immunologic reactions.
| Parameter | Typical range or observation | Practical implication |
|---|---|---|
| Systemic dose | ~1–2 × 10^13 vg/kg | High-dose requires immunologic monitoring |
| Packaging limit | ~4.7 kb single-stranded DNA | Limits cassette complexity |
| Second-strand synthesis | Rate-limiting, accelerated by helper functions | Expression can lag dose |
| Episomal persistence | Durable in non-dividing cells | Less durable in dividing tissue |
| Pre-existing immunity | Substantial in the human population | Can reduce efficacy |
| Liver clearance | Substantial fraction of systemically dosed vector | Reduces effective target dose |
| Anti-AAV antibodies | Prevalence varies by serotype and geography | Screening affects eligibility |
Timeline: from injection to durable effect
| Time after administration | What occurs | Clinical relevance |
|---|---|---|
| Minutes to hours | Capsid binds receptors and is internalized | Neutralizing antibodies can block this step |
| Hours to 1–2 days | Endosomal escape and cytoplasmic transit | Capsid degradation can reduce yield |
| 1–3 days | Nuclear entry and uncoating | Rate-limiting in post-mitotic cells |
| Days to 2 weeks | Second-strand synthesis and initial transcription | Transgene mRNA appears |
| 1–4 weeks | Protein expression rises toward plateau | When clinicians begin to look for effect |
| Weeks to months | Expression plateau or continued rise | Depends on promoter and target tissue |
| Months to years | Episomal persistence in non-dividing cells | Durable effect in neurons and hepatocytes |
| Years | Gradual loss in dividing tissues; rare integration | Long-term monitoring continues |
Edge cases and counterintuitive details
The most counterintuitive point is that a higher dose can be worse. At high systemic doses, capsid load can saturate clearance pathways and trigger complement activation, which is linked to thrombotic microangiopathy and other serious adverse events. The dose-response curve is not simply monotonic because immune activation rises faster than target-cell transduction at the high end.
Another counterintuitive detail is that the same capsid can transduce one tissue well and another poorly depending on the route. AAV9 given intravenously distributes broadly, but delivered directly to the cerebrospinal fluid it achieves very different central nervous system distribution. Route is part of the vector identity, not an afterthought.
A third point is that "non-integrating" is a term of degree, not an absolute. AAV DNA persists episomally but integration into the host genome occurs at low frequency. This is generally viewed as acceptable given the disease context, but it is why long-term follow-up is part of the regulatory expectation.
A professional caveat: neutralizing antibody titers are assay-dependent. A patient's eligibility can vary between laboratories and between serotypes. Two assays can give different "positive" results for the same sample, so an eligibility threshold quoted in one protocol may not transfer directly to another.
Misconceptions worth correcting
One common misconception is that AAV is engineered to integrate and cure by rewriting the genome. In most cases AAV provides a persistent episomal copy, and the durability depends on the target cell's division rate.
A second misconception is that one serotype fits every indication. Tropism is a combination of capsid, route, dose, and target tissue biology. A capsid that works in the retina may not work in the liver.
A third misconception is that higher doses always mean more expression. High doses raise immunologic risk and can saturate or overwhelm clearance.
A fourth misconception is that AAV is entirely inert immunologically. The capsid itself is immunogenic, and pre-existing immunity can block transduction. This is why antibody screening is standard in many programs.
A fifth misconception is that a small cassette is a minor design inconvenience. The ~4.7 kb limit forces trade-offs in promoter size, regulatory elements, and even the transgene itself. Some diseases need engineered mini-genes or dual-vector strategies because of this limit.
Practical actions for evaluating an AAV therapy
If you are assessing an AAV-based approach, the order of questions matters. Start with target tissue and route, then cassettes, then immune status.
Confirm the capsid and serotype match the target tissue and the administration route. Verify the therapeutic cassette fits within roughly 4.7 kb, including ITRs, promoter, and polyadenylation signal. Assess pre-existing anti-AAV neutralizing antibodies and understand the assay and threshold used in the relevant protocol. Review the dose per kilogram and the monitoring plan for liver, complement, and immune parameters. Confirm episomal persistence assumptions against the target cell's division rate, since expression durability differs in liver versus rapidly dividing tissue. Underline that this article is mechanism-level education, not a treatment plan. Clinicians set dosing and eligibility, and protocols differ by product and indication.
Limitations of the evidence and the platform
AAV transduction is well characterized at the cellular level, but several limits remain. Long-term durability beyond a few years is documented mainly in specific tissues such as the retina and liver, and extrapolating to all tissues is not supported by the evidence. Immunologic complications at high systemic doses are a known risk, and the exact thresholds for those risks are product-specific. Assay variability for neutralizing antibodies means eligibility criteria must be interpreted in the context of the specific program.
The platform's main engineering constraint, the packaging limit, is not a temporary inconvenience. It is a structural fact that shapes which diseases are approachable with a single vector. That is why the field invests in capsid engineering, promoter minimization, and delivery strategies such as dual vectors.
Finally, the durability question is inseparable from the tissue question. AAV transduction powers gene therapy most durably in slowly dividing or non-dividing cells, where the episome persists. In rapidly dividing tissue, the same vector will dilute with each division, and a durable cure may require repeat dosing or a different platform. That distinction is where much of the current clinical uncertainty sits, and it is why the field treats AAV as a strong but not universal tool.