Adeno-associated virus did not become the dominant in vivo gene therapy platform by accident or by consensus. It earned that position through a specific combination of biological properties that made it safer to administer, more predictable in its behavior, and more manufacturable at clinical scale than the alternatives available when the field was taking shape. Understanding why AAV succeeded where other platforms struggled — and where its genuine limitations lie — is essential context for anyone evaluating gene therapy programs today.
A Vector Born From a Biological Dependency
AAV’s origin as a gene therapy vector is rooted in a biological property that initially seemed like a liability: it cannot replicate on its own. AAV is a dependoparvovirus, meaning it requires co-infection by a helper virus — adenovirus or herpesvirus — to complete its replication cycle. Without helper functions, AAV integrates into the host genome at a specific locus on chromosome 19 (AAVS1) and goes dormant.
This dependency turned out to be a safety asset. A virus that cannot replicate autonomously poses a fundamentally lower risk of uncontrolled spread than replication-competent vectors. Early gene therapy programs using replication-competent adenoviral vectors encountered serious adverse events driven by uncontrolled viral replication and the immune responses it triggered. AAV, by contrast, could be engineered into a replication-incompetent vector — replacing its Rep and Cap coding sequences with the therapeutic transgene and supplying the packaging machinery separately — without sacrificing its ability to efficiently enter and transduce target cells.
The Non-Integrating Advantage
The most consequential safety distinction between AAV and the retroviral and lentiviral vectors that preceded it in clinical development is AAV’s predominantly non-integrating behavior. When recombinant AAV delivers its payload to the nucleus of a target cell, the single-stranded DNA genome is converted to a double-stranded form and exists primarily as episomal DNA — circular extrachromosomal elements that persist stably in non-dividing cells without inserting into the host chromosome.
This episomal persistence is significant for two reasons. First, it eliminates the risk of insertional mutagenesis — the disruption or aberrant activation of host genes at the site of integration — that was associated with early gamma-retroviral vector programs and contributed to leukemia cases in several pediatric patients treated for X-linked severe combined immunodeficiency. Second, it means the therapeutic gene is present in a predictable, non-heritable form that does not alter the germline.
The trade-off — that episomal DNA is diluted by cell division — is biologically real but clinically manageable for the applications where AAV performs best. Post-mitotic tissues such as neurons, photoreceptors, cardiomyocytes, and mature liver hepatocytes do not divide rapidly, and episomal AAV genomes are stable in these cells over years. The long-term expression data from approved AAV gene therapy programs — including decade-long follow-up in hemophilia B patients treated with AAV8-factor IX — has largely validated this stability assumption.
The Low Immunogenicity Profile
Compared to adenoviral vectors, AAV elicits a substantially lower innate immune response after administration. Adenoviral vectors are recognized by toll-like receptors and other pattern recognition molecules that trigger robust inflammatory signaling — a response that caused fatal systemic inflammation in at least one early clinical trial. AAV capsids do not contain the pathogen-associated molecular patterns that drive this response at the same intensity, and recombinant AAV preparations — when properly purified to remove empty capsids and process-related impurities — are generally well-tolerated at therapeutic doses.
This lower immunogenicity profile translated directly into clinical feasibility. It enabled systemic administration via intravenous infusion for diseases requiring body-wide distribution, such as spinal muscular atrophy, which would have been impractical with highly immunogenic vectors requiring immunosuppressive co-treatment at every administration.
Serotype Diversity as a Targeting Tool
The discovery and characterization of multiple AAV serotypes — each with a distinct capsid sequence and tissue tropism profile — gave the field a modular targeting toolkit that no other vector platform could match at the same safety level. By selecting the appropriate serotype for a given disease and route of administration, developers could bias vector biodistribution toward the liver (AAV8), the CNS (AAV9), the retina (AAV2, AAV8), or skeletal muscle (AAV1, AAV6) without modifying the therapeutic payload.
This modularity accelerated the translation of AAV from a single research tool into a platform capable of addressing a wide range of monogenic diseases. The same manufacturing infrastructure, the same regulatory framework, and the same preclinical safety assessment principles could be applied across programs differentiated primarily by their Cap sequence and transgene — enabling knowledge to accumulate across programs in a way that compounded over time.
Where AAV Genuinely Struggles
The biological properties that made AAV successful also define its limits, and those limits are not minor.
The payload constraint is the most discussed. The AAV capsid accommodates approximately 4.7 kilobases of single-stranded DNA. This is sufficient for a large number of therapeutically important genes, but it excludes others categorically. The full-length dystrophin gene — mutated in Duchenne muscular dystrophy, one of the largest markets in rare disease gene therapy — is approximately 11 kilobases in its coding sequence and cannot be packaged in a standard AAV capsid. Developers working in Duchenne have engineered truncated micro-dystrophin constructs that fit within the packaging limit, but the clinical performance of these shortened proteins relative to the full-length molecule remains under evaluation.
Re-dosing is a structural limitation with no clean solution. After a patient receives AAV gene therapy, the immune system mounts a response to the capsid proteins. Neutralizing antibodies generated against the initial capsid serotype will inactivate a second dose of the same serotype before it reaches target cells. Switching to a different serotype for re-dosing is theoretically possible but requires that the patient have no pre-existing immunity to the second serotype — a constraint that narrows options. For pediatric patients who receive AAV therapy before their immune systems have fully matured, or for patients whose therapy is delivered to an immune-privileged site such as the eye or CNS, re-dosing may be more feasible. For most adult systemic applications, single-dose efficacy is effectively a program requirement.
Pre-existing immunity in the general population further constrains patient eligibility. Natural exposure to wild-type AAV generates neutralizing antibodies in a significant proportion of adults — with seroprevalence varying by serotype and geography but reaching 40 to 70 percent for AAV2 in some populations. Clinical trials for AAV gene therapies routinely screen patients for pre-existing antibodies and exclude those above a defined titer threshold, which reduces the addressable trial population and raises questions about commercial market size.
Manufacturing cost and scale remain unresolved challenges for the field. Most approved AAV gene therapies command prices in the range of one to three million dollars per dose — pricing that reflects, in part, the cost of manufacturing a single-dose biological product at small batch sizes with extensive analytical characterization requirements. Scale-up has proven more difficult for AAV than for conventional biologics, and the industry continues to grapple with the economics of making gene therapy commercially sustainable beyond rare orphan disease indications.
The Platform in Context
AAV’s dominance in in vivo gene therapy is deserved by the evidence. Its safety record, biodistribution predictability, long-term expression stability, and serotype versatility have enabled the approval of treatments for diseases that had no effective options. But the platform is not universal, and the programs that have encountered the most difficulty are those that treated AAV as a default rather than as a considered choice among alternatives with different profiles. The payload limit, the re-dosing constraint, the immunogenicity landscape, and the manufacturing economics are not minor footnotes — they are design parameters that should be incorporated into program strategy from the beginning.