Design It Right: Evaluating AAV Manufacturability Through Molecular and Upstream Optimization
This article was first published in Endpoints News.
Author Contributing Authors
The rapid expansion of gene therapy pipelines has made scalable and reliable adeno-associated virus (AAV) production increasingly important. Moving from early proof-of-concept to a cGMP-ready process requires more than scaling volumes—it requires attention to how vector design and production systems interact. Variables such as plasmid architecture, transgene expression, and vector genome stability can determine whether a program meets yield and quality goals. Considerations like serotype-specific performance, transfection conditions, and host cell line compatibility are equally critical, as these factors directly affect both productivity and regulatory readiness.
At Forge Biologics, manufacturability is built into every program from the very beginning. By evaluating vectors at the molecular level and through structured process development studies, clients gain early insight into how constructs are likely to perform and where adjustments may strengthen outcomes. Forge’s proactive approach enables smoother transitions to cGMP production and supports long-term scalability, potentially reducing the risk of additional time and costs with rework later in clinical development.
Plasmid Assessment as a Manufacturability Tool
Forge offers a detailed plasmid evaluation for developers seeking to de-risk manufacturability at the sequence level before committing to process development scale-up work. This molecular development package provides insights into potential bottlenecks and highlights elements that may influence productivity, stability, or safety. In addition, developers benefit from the extensive vector design experience of Forge’s technical experts, led by Chief Technology Officer David Dismuke, Ph.D., and Linas Padegimas, Ph.D., Senior Director of Research and Development. They bring over four decades of experience in vector design consultation to each project.
This evaluation typically considers inverted terminal repeat (ITR) size and integrity, homology between gene of interest (GOI) and rep/cap regions, regulatory element architecture, open reading frame (ORF) continuity, untranslated regions (UTR) structure, and overall cassette size relative to AAV packaging constraints. Even small differences in these features can alter genome packaging efficiency, transgene expression levels, or stability once the vector is in production1-3 (Figure 1).
One example illustrates this clearly (Figure 2). In this case, the original AAV2 rep/cap plasmid design contained flanking homology to the GOI plasmid. The rep/cap plasmid was switched to Forge’s redesigned, modified FUELTM rep/cap plasmid, eliminating the overlapping sequences while maintaining the AAV2 capsid. This adjustment not only resolved the risk of replication-competent AAV (rcAAV) formation but also increased productivity up to five-fold under the same conditions, translating into a meaningful improvement to output.
When vectors are built on a strong molecular foundation, with well-structured, functional design and minimal recombinogenic elements, they are more likely to scale successfully into large-scale production. In contrast, sequence liabilities can create challenges that downstream optimization alone cannot fully address. 
Functional Screening: From Design to Demonstrated Performance
Plasmid assessments offer valuable clues about how a construct might perform, but manufacturability only becomes clear when those designs are tested under production-like conditions. At Forge, this step involves small-scale shake flasks or Ambr® 250-based screening, using Forge’s HEK293 Ignition™ cell line and a standard triple transfection process. Each candidate is run under matched conditions so that differences in transfection efficiency, titer, and cell viability can be observed without confounding variables.
This stage often shows that constructs with similar design parameters can behave very differently in practice (Figure 3A). Some consistently deliver high titers, while others show limitations only evident under scalable conditions. Early functional screening helps identify constructs with the strongest potential for scale-up, while avoiding investment in designs with inherent productivity or packaging limits. In this way, screening bridges sequence-level hypotheses with the practical realities of vector production. 
Integrating DOE, Downstream Evaluation, and Quality
Once a lead candidate or a small set of candidates has been identified, the next step is to refine performance using design of experiments (DOE). Unlike one-variable testing, DOE explores multiple factors in combination, such as plasmid ratios, DNA concentration, transfection reagents, and use of enhancers to capture interactions and identify the conditions that drive the most consistent results.4 In practice, blending careful vector design with DOE-guided optimization has led to stepwise gains, with overall yields improving more than 20-fold in some cases (Figures 3B-3D).
It is equally important to extend testing beyond upstream production. Candidates are advanced into downstream purification and characterization, where factors such as full-to-empty ratios, step recoveries across purification operations, and impurity clearance profiles are evaluated. These assessments provide a fuller picture of manufacturability and ensure that construct selection is guided not only by upstream productivity but also by downstream compatibility.
Downstream evaluation also allows for the measurement of critical quality attributes (CQAs) once drug product material is available. Attributes such as rcAAV risk, capsid integrity, genome integrity, impurity levels, and potency can be examined to begin shaping an early target product profile. This material is also valuable for developing or confirming assays still in progress—such as potency or infectivity testing—ensuring that analytical methods advance alongside process development. Collecting these data, even at an initial stage, provides important context for construct choice and establishes a foundation for later-stage development.
When coupled with platform-specific enhancements such as Forge’s FUELTM system, the impact of DOE becomes even clearer: optimized clinical candidates have demonstrated 6-7x productivity gains compared to industry standard, with further DOE-driven transfection optimization amplifying yields up to 22x (Figure 4). Integrating these upstream productivity gains with downstream evaluation ensures that manufacturability is addressed holistically, linking higher titers to quality and compatibility needed for confident construct selection and process lock. 
Considerations Before Tech Transfer
Programs approach tech transfer at different levels of readiness, and Forge is structured to support both. Some developers arrive with primarily upstream data, which provides a first look at vector performance but leaves downstream behavior less defined. In these cases, Forge can partner to design and execute the additional purification and analytical studies needed to evaluate capsid integrity, full-to-empty ratios, and impurity clearance profiles before finalizing a construct for cGMP production.
Other developers come with a more comprehensive package: upstream and downstream evaluations completed, and a candidate already selected. These projects may be positioned for a faster transition but still benefit from Forge’s ability to align prior work with platform processes, confirm reproducibility at scale, and close any remaining analytical or regulatory gaps. Forge offers a full suite of regulatory services to support programs at any stage.
In either scenario, there are common considerations that increase the likelihood of success before tech transfer. Developers should aim to characterize plasmid maps with attention to regulatory and homology elements, confirm cassette size, assess transgene expression, and gather at least preliminary data on both productivity and critical quality attributes. The more of this foundation that is in place, the smoother and more efficient the path into CDMO-led process development.
Conclusion
Manufacturability is best understood as the intersection of molecular design and process responsiveness. By evaluating vectors at the sequence level, validating them under scalable conditions, and applying DOE to refine parameters, Forge Biologics has developed a structured approach that supports more predictable outcomes. This strategy not only helps clients avoid costly setbacks but also builds confidence that selected constructs can advance efficiently into cGMP production.
At its core, manufacturability can be viewed as a continuum from sequence to scale. This perspective helps clients anticipate challenges, make confident choices, and position their programs for the demands of clinical and commercial manufacturing, ensuring smoother development pathways. Forge’s FUEL™ platform, with its purpose-built cGMP facility, and a team experienced in both scientific and regulatory complexities, provides partners with a unique advantage: the ability to link early molecular design with large-scale manufacturability. This holistic approach ensures that programs are not only scientifically sound but also commercially viable, paving the way for more reliable gene therapy development and ultimately, greater impact for patients.
References
- Blahetek G., et al. AAV yield, bioactivity, and particle heterogeneity are impacted by genome size and non-coding DNA elements. Molecular Therapy Methods & Clinical Development, 2025 Jun 2;33(3):101499. doi: 10.1016/j.omtm.2025.101499. PMID: 40590037; PMCID: PMC12207685.
- Allen J.M. et al. Identification and elimination of replication-competent adeno-associated virus (AAV) that can arise by nonhomologous recombination during AAV vector production. J. Virol. 1997; 71:6816–6822. doi: 10.1128/jvi.71.9.6816-6822.1997.
- Asaad W., et al. AAV genome modification for efficient AAV production, Heliyon, Volume 9, Issue 4, 2023, e15071, ISSN 2405-8440, https://doi.org/10.1016/j.heliyon.2023.e15071.
- Tzimou, Konstantina et al. Unlocking DOE potential by selecting the most appropriate design for rAAV optimization, Molecular Therapy Methods & Clinical Development, Volume 32, Issue 4, 101329.
