Among the quality attributes measured in AAV gene therapy manufacturing, the ratio of full to empty capsids is one of the most consequential and least intuitively understood. It is not a proxy for something else. It is a direct determinant of product potency, patient safety, and manufacturing economics — and it has become a central point of focus in regulatory review of AAV gene therapy submissions. Understanding what full and empty capsids are, why the distinction matters, and what can be done about it requires working through the biology and the manufacturing logic together.

What Is an AAV Capsid?

The AAV capsid is the protein shell that surrounds and protects the viral genome. It is composed of 60 viral protein (VP) subunits — VP1, VP2, and VP3 — arranged in an icosahedral geometry approximately 25 nanometers in diameter. The capsid serves multiple functions beyond simple containment: it mediates binding to cell surface receptors, facilitates endosomal escape after cellular uptake, enables nuclear localization, and is the primary target of the host immune response.

In the manufacturing process, capsid assembly and genome packaging are not perfectly synchronized. The VP proteins assemble into capsid shells continuously during the production window, while the Rep proteins work in parallel to replicate vector genomes and thread them into assembling capsids. When these two processes are not in balance — when capsids assemble faster than genomes are packaged — the result is empty capsids: fully assembled protein shells containing no therapeutic DNA.

The Three Populations

AAV preparations produced by transient transfection or other manufacturing methods contain a mixture of particle populations that can be distinguished by their genome content.

Full capsids contain a complete, packaged vector genome — a single-stranded DNA molecule of approximately 4.7 kilobases flanked by inverted terminal repeats. These are the intended product: the particles that will deliver the therapeutic gene to target cells after administration. Full capsid content is the most direct measure of vector potency, and the viral genome titer — expressed as viral genomes per milliliter — quantifies the number of full capsids in a preparation.

Empty capsids are assembled protein shells containing no DNA payload. They are structurally indistinguishable from full capsids by most visual assays, and they are present in every AAV manufacturing process to some degree. In poorly optimized processes, empty capsids can constitute the majority of the total particle count.

Partial capsids contain fragmented or truncated genomes — DNA sequences shorter than the full vector genome, often representing packaging of truncated or rearranged sequences generated during genome replication. Partial genomes may lack complete expression cassettes and therefore do not contribute to therapeutic activity. They may also complicate identity testing and potency measurement if they are present at significant levels.

Why Empty Capsids Matter Clinically

Empty capsids are not inert. They are antigenically identical to full capsids — their surface protein sequences are the same — which means the immune system treats them identically. Every empty capsid administered to a patient generates the same capsid-directed immune response as a full capsid, without contributing any therapeutic benefit.

This has direct implications for dosing and safety. The total dose of AAV administered to a patient is typically described in terms of viral genomes — the number of full capsids — but the total particle burden experienced by the patient’s immune system includes all capsids, full and empty. A preparation with a high empty capsid content requires a higher total particle dose to deliver the same number of viral genomes, increasing the immunological burden without increasing the therapeutic payload.

At the doses used in systemic gene therapy for diseases such as Duchenne muscular dystrophy or hemophilia — doses in the range of 10¹³ to 10¹⁴ viral genomes per kilogram of body weight — this is not a theoretical concern. High-dose AAV gene therapy has been associated with serious adverse events including complement activation, thrombotic microangiopathy, and acute liver injury. The contribution of empty capsids to these responses, relative to full capsids, is an area of active investigation, but the regulatory principle is clear: minimizing unnecessary particle burden is a manufacturing quality objective with direct patient safety implications.

Why Regulators Care

The FDA’s expectation that AAV sponsors characterize and control the full/empty capsid ratio in their drug product is grounded in the safety and potency logic above. The ratio is considered a critical quality attribute — a product characteristic that must be within a defined range to ensure the desired safety and efficacy profile.

Characterization of the full/empty ratio is required for IND submissions and becomes progressively more rigorous as programs advance toward BLA. Sponsors must demonstrate not only that they can measure the ratio accurately and reproducibly, but that their manufacturing process produces a consistent ratio across batches, and that their downstream purification process enriches full capsids to the required specification.

Comparability of full/empty ratios across manufacturing changes — scale-up, process optimization, site transfer — is a standard regulatory question. A change in manufacturing that alters the full/empty ratio without adequate comparability data is treated as a change that may affect product quality and potency, requiring characterization and potentially re-submission of supporting data.

Lot release specifications for commercial AAV products typically include an upper limit on empty capsid content or a lower limit on the percentage of full capsids. These specifications must be justified with clinical and analytical data and are subject to regulatory review and approval.

Analytical Methods for Full/Empty Characterization

Several analytical methods are used to characterize AAV full/empty capsid content, each with different strengths and limitations.

Analytical ultracentrifugation (AUC) separates full and empty capsids based on their sedimentation coefficients, which differ because of the mass difference between a DNA-containing and a DNA-free particle. AUC is considered the reference method for full/empty determination and is widely accepted by regulators, but it is low-throughput and requires specialized instrumentation.

Transmission electron microscopy (TEM) allows direct visual counting of full and empty capsids, distinguishing them by the electron density of the particle interior — full capsids appear dark due to the DNA content, while empty capsids appear light. TEM is informative but low-throughput, operator-dependent, and not well-suited to routine lot release.

Charge detection mass spectrometry (CDMS) and other emerging mass-based methods can distinguish full, partial, and empty populations with high resolution and are increasingly used for process development and characterization, though regulatory acceptance as a lot release method continues to develop.

The ratio of total particle titer (measured by ELISA or similar immunoassay detecting VP proteins) to genome titer (measured by droplet digital PCR or quantitative PCR) provides an indirect estimate of the full capsid fraction. This approach is widely used because both components are already measured for other purposes, but it is less precise than direct separation methods and more sensitive to assay-specific biases.

Comparison of analytical methods for AAV full/empty capsid characterization. AUC = Analytical Ultracentrifugation; TEM = Transmission Electron Microscopy; CDMS = Charge Detection Mass Spectrometry; ddPCR = Droplet Digital PCR. Status reflects current regulatory acceptance and industry adoption as of 2024. Established methods are accepted for lot release; Emerging methods are increasingly cited in submissions but regulatory acceptance is still developing.

Manufacturing Strategies to Improve the Ratio

The full/empty capsid ratio is not fixed by the biology of AAV — it is substantially influenced by manufacturing choices.

At the upstream level, the balance between Rep and Cap expression affects packaging efficiency. Rep protein levels determine how efficiently vector genomes are replicated and threaded into assembling capsids. Optimizing plasmid ratios during transfection — typically adjusting the relative amounts of Rep-Cap plasmid, transfer plasmid, and helper plasmid — can shift the balance toward more productive genome packaging and reduce empty capsid generation.

Transfection conditions including cell density, transfection reagent, DNA-to-reagent ratios, and timing of harvest affect both total vector yield and the full/empty distribution of that yield. Process development specifically aimed at improving the full capsid fraction is now considered a standard component of AAV manufacturing optimization.

At the downstream level, purification strategies can enrich full capsids by exploiting the density difference between full and empty particles. Cesium chloride density gradient ultracentrifugation — historically the standard purification approach — effectively separates full from empty capsids based on buoyant density. Modern chromatographic purification platforms, while generally preferred for scalability and GMP compliance, may require additional steps such as ion exchange chromatography or size exclusion to achieve comparable full capsid enrichment.

The downstream full capsid enrichment step is not merely a quality polishing step — it directly affects the potency per total particle of the final drug product and the total particle burden delivered to each patient. Programs that treat it as an afterthought in process design routinely encounter lot release challenges and patient safety questions that upstream decisions did not anticipate.

Full/empty capsid characterization and control sits at the intersection of analytical science, process development, and regulatory strategy. Getting it right requires treating it as a design priority rather than a characterization exercise conducted after the manufacturing process is already set.