Among the decisions made in early AAV gene therapy development, serotype selection is among the most consequential and least forgiving. The serotype of an AAV vector determines which tissues it preferentially infects, which cell surface receptors it uses for entry, how efficiently it transduces the target cell population, and how the immune system responds to the capsid. Changing the serotype late in development — after safety studies, process development, and regulatory interactions have been built around a specific capsid — is not a course correction. It is effectively a program restart.
Understanding tissue tropism, what drives it, and how serotype selection fits into a coherent development strategy is therefore not a background scientific detail. It is a front-end strategic decision with long-term program consequences.
What Tissue Tropism Means
Tropism refers to the tendency of a virus — or a viral vector — to preferentially infect certain cell types over others. For AAV, tropism is primarily determined by the capsid protein sequences that mediate cell surface binding and receptor-dependent uptake. Different serotypes interact with different cell surface receptors, and those receptor distributions across tissues determine where the vector goes and how efficiently it transduces cells at each location.
Tropism is not binary. A serotype that is described as liver-tropic does not exclusively transduce hepatocytes — it transduces hepatocytes most efficiently relative to other cell types under typical administration conditions. Other tissues may also be transduced, particularly at high doses, which has safety implications that must be characterized during preclinical development and monitored in clinical trials.
The Biology Behind Serotype Differences
AAV serotypes — numbered AAV1 through AAV13, plus a growing catalog of engineered and naturally occurring variants — differ in the amino acid sequences of their VP1, VP2, and VP3 capsid proteins, particularly in the variable surface loops that protrude from the icosahedral capsid surface. These loops mediate the initial interaction with cell surface receptors and co-receptors and are the primary molecular basis for tropism differences between serotypes.
The primary receptor for AAV2 — the first serotype to be extensively characterized — is heparan sulfate proteoglycan (HSPG), a ubiquitous cell surface molecule present on many cell types. AAV2 was consequently one of the first serotypes to enter clinical development and produced transduction across a relatively broad range of tissues, though with particular efficiency in retinal cells, liver, and muscle under relevant administration conditions.
AAV8 uses the laminin receptor and members of the LamR family as its primary entry receptor, and this receptor profile correlates with its strong hepatocyte tropism. AAV8 transduces the liver with substantially higher efficiency than AAV2 following intravenous administration and at lower doses, which is why it has been the serotype backbone for most hemophilia gene therapy programs. AAV5, which uses platelet-derived growth factor receptor as its primary receptor, transduces muscle and liver with reasonable efficiency and has been used in hemophilia A programs with factor VIII payloads.
AAV9 has the distinctive ability to cross the blood-brain barrier after systemic intravenous administration — a property associated with its use of galactose as a primary receptor, which is expressed on endothelial cells lining the cerebral vasculature. This property made AAV9 the capsid of choice for Zolgensma, the SMN1 gene therapy for spinal muscular atrophy, administered as a single intravenous infusion that achieves widespread CNS transduction in infants. AAVrh10, a non-human primate-derived serotype with strong CNS penetration, has been used in several neurological disease programs as an alternative to AAV9 in contexts where the immune landscape differs.
AAVPh.B1 and other engineered capsids developed through directed evolution and rational design approaches have been specifically optimized to enhance transduction of particular target tissues — including photoreceptors, cardiomyocytes, and CNS neurons — while reducing off-target uptake in liver. This class of engineered capsids is increasingly entering clinical development and represents the next generation of serotype-specific optimization.
The Role of Route of Administration
Tropism is not determined by serotype alone. The route of administration is a co-determinant of tissue targeting and cannot be optimized independently of capsid selection.
Intravenous administration exposes the vector to the entire vascular compartment, and the capsid must navigate liver sinusoids — where Kupffer cells actively sequester foreign particles — before reaching other tissues. Serotypes with strong liver tropism such as AAV8 and AAV9 are efficiently taken up by hepatocytes after IV dosing. Serotypes that must reach muscle, CNS, or other tissues via IV administration face higher doses and the immune exposure that comes with systemic circulation.
Intramuscular injection deposits the vector locally, relying on muscle fiber uptake and local transduction rather than vascular distribution. Serotypes with muscle tropism — AAV1, AAV6, and AAV8 under certain conditions — are used in intramuscular programs, and this route avoids the hepatic first-pass effect that complicates IV dosing for some applications.
Intravitreal and subretinal injections are used for ocular gene therapy. Subretinal injection, used in Luxturna (AAV2-RPE65), deposits the vector in direct contact with the retinal pigment epithelium and photoreceptor outer segments, effectively bypassing tropism as a primary targeting mechanism and relying instead on proximity. Intravitreal injection requires a capsid capable of penetrating the vitreous and diffusing to the appropriate retinal cell layer — a traversal that favors smaller capsids and specific surface charge properties.
Intrathecal administration, used for some CNS programs, deposits the vector in the cerebrospinal fluid, allowing distribution along the neuraxis to reach motor neurons and other CNS cell types. This route achieves broader CNS distribution than parenchymal injection at lower total doses compared to IV.
Pre-existing Immunity and Serotype Selection
A practical constraint on serotype selection that is often underweighted in early development is the prevalence of pre-existing neutralizing antibodies to specific serotypes in the general population. Wild-type AAV serotypes circulate in human populations, and prior natural exposure generates seroneutralizing antibody responses that can render patients ineligible for clinical trials or, in approved products, require pre-screening and exclusion of seropositive patients.
AAV2 has the highest seroprevalence of the commonly used serotypes — with neutralizing antibodies detected in 40 to 70 percent of adults in some studies, depending on the population and assay. AAV1, AAV6, and AAV9 have somewhat lower but still clinically meaningful seroprevalence. AAV8 has among the lowest seroprevalence of the commonly used serotypes, which is one factor contributing to its dominance in liver-directed programs. Some engineered capsids have been specifically designed to reduce binding by pre-existing antibodies while retaining target tissue tropism.
The seroprevalence landscape has direct implications for clinical trial eligibility criteria, addressable patient populations, and commercial market size. A serotype selection that is optimal for transduction efficiency but excludes 50 percent of the target patient population due to pre-existing immunity creates a very different commercial ceiling than one with lower seroprevalence.
Serotype Selection as Program Strategy
The decision tree for serotype selection in practice requires integrating several factors simultaneously: the target tissue and cell type, the required dose and transduction efficiency, the route of administration, the payload size and its compatibility with the capsid, the seroprevalence of candidate serotypes in the target patient population, and the prior regulatory experience with the candidate capsid.
For diseases with well-established serotype precedents — liver-directed factor replacement for hemophilia, CNS delivery for neurological diseases — the serotype choice is informed by a body of clinical data that reduces scientific risk. Programs that deviate from established capsid precedents — using engineered capsids or non-standard serotypes — may achieve superior biology but carry greater regulatory uncertainty and require more extensive preclinical characterization to support IND applications.
This is where serotype selection becomes a strategic decision rather than purely a scientific one. The capsid that provides the best transduction efficiency in a mouse model or a non-human primate is not necessarily the capsid that provides the fastest, lowest-risk path through clinical development. Aligning transduction biology, regulatory precedent, manufacturing capability, and patient eligibility criteria requires a front-end analytical investment that pays disproportionate returns in development timeline and capital efficiency.
Programs that treat serotype selection as a revisable early-stage assumption — something that can be changed once the transgene is validated — routinely discover that the capsid is not separable from the program in the way that assumption implies. Every preclinical safety study, every manufacturing process, every regulatory interaction is capsid-specific. Serotype selection is the architecture on which the rest of the development program is built.
MKA Insights works with CGT developers at the early stages of program design to integrate serotype selection into a coherent development and regulatory strategy. If your program is at the stage where vector platform and capsid decisions are being made, we bring the analytical rigor and cross-domain perspective that these foundational choices require.
MKA Insights