Podcast recap: David Dismuke and Steven Gray on scaling AAV gene therapy

AAV gene therapy has moved from proof of concept to approved products. As more programs enter the clinic and target larger patient populations, manufacturing has become a defining factor in what is feasible. Process design, vector architecture, analytics, and scale all shape how quickly therapies reach patients and at what cost.

In the last episode of The Genetics Podcast, Patrick Short spoke with Dr. David Dismuke, CTO at Forge Biologics, and Dr. Steven Gray, Professor and Co-Director of the Gene Therapy Program at UT Southwestern. Their conversation connects academic vector development with industrial scale production and highlights how delivery efficiency, process productivity, and product quality intersect.

From academic vector cores to industrial platforms

Both guests trained in Jude Samulski’s lab and saw early on that manufacturing would determine whether AAV therapies could move beyond small, investigator-led studies.

Forge Biologics was built around a platform model for AAV production. The goal is to avoid rebuilding the process for every new client program. Instead of starting from scratch each time, companies can plug their transgene into an established manufacturing backbone and focus on the biology.

At a high level, the manufacturing process involves:

• Expanding HEK293 cells in single use bioreactors
• Transiently transfecting cells with DNA encoding capsid components and the therapeutic payload
• Harvesting after several days
• Purifying through affinity chromatography and polishing steps

Product heterogeneity is a central challenge

AAV manufacturing does not generate a single uniform product. Each batch contains a mixture of particles, including fully packaged capsids containing the therapeutic gene, empty capsids, and partially packaged capsids.

Clinical dosing is typically calculated based on full particles. However, total capsid exposure matters. Empty particles increase the overall antigen load without increasing therapeutic benefit, which can influence immune responses at higher doses.

Crucially, product quality is shaped by the interaction between vector design and manufacturing. AAV has a packaging limit of roughly 4.7 kilobases. What goes into the capsid directly affects how efficiently it can be produced.

Downstream purification can enrich for full capsids, but tighter separation usually reduces yield. That tradeoff forces teams to think about manufacturability early. Vector design decisions influence process efficiency, cost, and safety margins, not just biological performance.

Analytics and the clinical feedback loop

As programs advance into the clinic, manufacturing attributes are no longer abstract metrics. They become part of the clinical story.

When adverse events occur, developers revisit:

• Dose levels
• Vector design
• Empty to full ratios
• Total capsid load
• Potency assays

Similarly, if efficacy is weaker than expected, teams may reevaluate promoter strength, tissue targeting, expression control elements, or delivery efficiency assumptions.

Analytical development is therefore a major component of AAV programs. Quantifying vector genomes, characterizing particle populations, and establishing functional potency assays require time and regulatory rigor. For licensure, potency assays often need to reflect biological function rather than simple expression, adding another layer of complexity.

Manufacturing and clinical development operate in a feedback loop. Process data informs clinical interpretation, and clinical outcomes drive process refinement.

Cost, productivity, and the path to broader adoption

The cost of AAV therapies reflects several factors. Development risk is high, clinical trials are expensive, and manufacturing requires specialized infrastructure, skilled teams, and high quality raw materials.

At the same time, the field has made measurable productivity gains. Over the past decade, vector yields per liter have improved substantially. Higher titers translate into smaller required batch sizes for the same clinical supply. A program that once required very large bioreactor volumes may now be feasible at more modest scales.

Delivery improvements also influence cost. As Steven notes, gene therapy is fundamentally about delivery. If capsids become significantly more efficient at targeting specific tissues, the required dose can decrease. Lower dose requirements reduce manufacturing demand and can widen safety margins.

Recent advances in engineered capsids illustrate this trend. Instead of relying exclusively on naturally occurring serotypes, researchers are designing capsids with improved tissue targeting, including variants that more effectively cross the blood brain barrier. Greater efficiency can unlock new applications and make previously impractical strategies, such as complex gene editing payloads, more viable.

Rare diseases and common diseases may require different models

As AAV programs expand into more common conditions, manufacturing strategy may need to evolve.

Current facilities often emphasize flexibility. Single use systems and modular clean rooms allow rapid changeover between programs, which suits rare disease portfolios with diverse vectors and limited batch sizes.

Larger indications introduce different requirements. Supplying hundreds of thousands or millions of patients may require dedicated facilities, larger scale bioreactors, greater process standardization, and potential shifts toward producer cell lines rather than transient transfection.

These are distinct engineering problems. One focuses on flexibility and rapid iteration across many programs. The other prioritizes sustained, high volume output for a single product. The field may ultimately require both models operating in parallel.

Looking ahead

Both guests are optimistic about the next phase of AAV. Traditional gene replacement remains important, particularly for monogenic diseases. At the same time, improved delivery platforms could support gene editing approaches such as base or prime editing, provided efficiency continues to improve.

Manufacturing scale, delivery efficiency, and product quality are not secondary considerations. They determine whether promising science can translate into durable, widely accessible therapies. For AAV to move from specialized treatments toward broader clinical impact, advances in vector engineering and manufacturing science will need to progress together.

Why this matters for trial sponsors

Manufacturing realities directly inform clinical strategy in AAV programs. Dose selection is influenced not only by biological rationale but also by what can be produced consistently and administered within acceptable safety margins. If a therapy requires high systemic doses that approach immunological thresholds or strain manufacturing capacity, sponsors may need to narrow eligibility criteria, adjust weight bands, or adopt more conservative dose escalation schemes. Clinical design decisions are often intertwined with supply constraints and product characteristics.

Yield variability also affects development planning. When purification reduces recoverable material or batch output is inconsistent, sponsors must factor supply risk into cohort sizing and enrollment timelines. Expansion decisions and staggered dosing strategies frequently reflect confidence in manufacturing reproducibility as much as emerging efficacy data.

Product attributes such as total capsid load and empty to full ratios influence safety monitoring intensity and protocol conservatism. These considerations can shape everything from inclusion criteria to the structure of interim analyses.

Finally, scalability influences long term indication strategy. A program that is feasible for a narrowly defined rare disease population may require meaningful process optimization before expanding into broader or earlier stage patient groups.

Listen to the full episode below.