Rapidly growing interest in gene therapy has led to the need for more cost-effective and scalable viral-vector manufacturing platforms. Adenoassociated virus (AAV) has become a vector of choice because of its safety profile (nonpathogenic infection). In addition, AAV cannot replicate on its own and is not integrated directly into the host genome.

AAV vector manufacturing using human embryonic kidney (HEK) cells in either adherent or suspension mode includes several typical processing steps: cell expansion, plasmid transfection, viral-vector production, cell lysis, purification, and fill and finish. The purification process usually involves clarification, capture through affinity chromatography, polishing with anion-exchange (AEX) chromatography, tangential-flow filtration (TFF) concentration/diafiltration, and final filtration. An additional TFF process step often is added before the affinity step to concentrate the product stream.

Those upstream and downstream processes can be fairly straightforward at small scales. However, obstacles often are encountered when a developer progresses toward larger volumes for clinical and commercial manufacturing. In particular, obstacles often come up during transfection, clarification, and the separation of empty and full capsids.

Overcoming Transfection Challenges
The first major obstacle to scaling up AAV manufacturing processes arises in the transfection step. Whether cells are adherent or suspension-adapted, their transfection at large scales brings a unique set of challenges. Overall, a controlled and scalable process must be established to ensure a robust process.

First, traditional chemical methods — e.g., based on calcium phosphate — work well at small scales but rapidly become limited when scaled up. Raw materials used must comply with good manufacturing practices (GMPs) for large-scale manufacturing and be readily available. Thus, transfection reagents such as polyethylenimine (PEI) are preferred.

Reagent manufacturers are developing a series of compounds specifically for large-scale transfection processes. The new alternatives are optimized for specific types of viral vectors, depending on whether the process uses adherent or suspension-adapted cells. Some newly released transfection reagents dedicated specifically to AAV vector production in suspension-based systems could provide two to three times higher yields than commonly used generic PEI reagents can give.

Second, because of the great volumes of plasmid DNA (pDNA) and PEI required for large-scale manufacturing, time must be allotted for creating the necessary pDNA–PEI complexes and transferring them to a bioreactor. The sizes of those complexes also must be controlled to ensure optimal transfection efficiency. Low-shear pumps must be used to prevent damage to the complexes during transfer.

Finally, using a scalable bioreactor platform greatly facilitates the transfer of such processes from development laboratories to manufacturing plants. For instance, Pall offers the Allegro STR bioreactor platform from 50 L to 2,000 L for suspension processes.

For adherent processes Pall offers iCELLis fixed-bed technology. It mimics the conditions of flatware for easy transfer of initial processes to a bioreactor environment. Rapid scale-up follows from an iCELLis Nano bioreactor to an iCELLis 500+ bioreactor (manufacturing-scale model). The technology has been widely accepted in the biopharmaceutical industry. For example, iCELLis fixed-bed bioreactors are used in production of Zolgensma (onasemnogene abeparvovec-xioi), an AAV-based gene therapy used to treat patients with spinal muscular atrophy.

Clearing the Clarification Hurdle
Clarification is a critical step in AAV vector manufacturing because culture supernatants produced by transfection are contaminated by large amounts of pDNA, and cell lysis generates a significant quantity of host-cell contaminants (both proteins and DNA). Clarification is a necessity and must proceed with high throughput for impurity removal, high product yield, and ease of scale-up.

The initial removal of cell debris and impurities during clarification typically involves a combination of filters with decreasing pore sizes down to 0.2 μm. Including a prefilter with relatively large nominal pore sizes is essential to removing cellular debris and protecting subsequent filters from fouling. In selection of appropriate filters, users should take into consideration throughput, turbidity reduction, volume that can be processed in the target filter capsule size, product yield, and overall economics.

In one study conducted to identify an optimized clarification process for a suspension AAV feed, the authors considered depth filters, prefilters, and bioburden reduction membrane filters (1). Two solutions were found to serve as viable methods: the dual-layer, single-step of Seitz HP PDH11 or Seitz HP PDK11 filters and the triple-layer, dual-step combination of a Seitz HP PDP8 filter configured in series with a Seitz Bio 10 filter. Overall, the latter combination exhibited the highest throughput, the highest yield, and potentially the most economical option.

Separating Empty and Full Capsids
Typically, affinity chromatography serves as an initial bulk “cleanup” step that will be followed by AEX chromatography in a polishing step. The biggest challenge during AEX purification is to remove empty virus capsids while retaining full capsids that contain the complete genetic payload. Questions still remain about the value of those empty capsids, but regulatory authorities consider them to be in-process contaminants and therefore expect the percentage of empty capsids to be minimized and the empty/full ratio to be controlled at all times to ensure gene-therapy product consistency.

Achieving that goal can be challenging because the empty/full ratio resulting from AAV upstream processes from current production technologies often varies from batch to batch. Ultracentrifugation is a good tool for measuring that ratio at the laboratory scale by separating empty from full capsids. But it is impractical at manufacturing scales. AEX chromatography thus has become the most widely used method for this purpose.
In addition to the multiple natural serotypes of AAVs, hybrid serotypes designed to target particular sites in a patient’s body are being developed. Each of those serotypes behaves differently during chromatographic separations, making it difficult for developers to create platform purification processes.

Traditional chromatographic approaches used for separation of empty and full capsids are based on monoliths or columns packed with resin. A promising alternative is the use of membrane adsorbers. Used with small conductivity-step changes (rather than linear gradients), Pall Mustang Q membranes can yield distinct elution peaks for both DNA-free and DNA-containing capsids.

Help Is on the Way
As the gene therapy market grows and more candidate products advance into late-stage clinical trials and through to commercialization, many new AAV and other viral-vector processes must be scaled to larger volumes. Process developers face several challenges to achieving highly efficient, productive, and robust solutions, including the issues with transfection, clarification, and chromatography described herein.

Some of those challenges ultimately will be addressed with greater understanding of relevant biology and advances in upstream manufacturing technologies. For instance, development of stable, high-producing cell lines will eliminate the need for transfection and could provide consistency in empty/full capsid ratios.

In the meantime, gene-therapy developers must leverage existing downstream purification technologies to optimize manufacturing processes. Overall yields across downstream purification steps from clarification through final filtration can be low, leaving room for future optimization.

Pall provides end-to-end solutions for gene-therapy manufacturing, including the Accelerator process development services. The unique combination of services, novel technologies, and proven methods can be leveraged to develop processes, optimize yields, and scale customized versions of the resulting platform technologies to meet the specific needs of individual gene-therapy projects.

Reference
1 Raghavan B, et al. Optimizing the Clarification of Industrial Scale Viral Vector Culture for Gene Therapy. Cell Gene Ther. Ins. 5(9) 2019: 1311–1322; https://doi.org/10.18609/cgti.2019.137.

Emmanuelle Cameau is viral vector and gene therapy technical manager at Pall Corporation, Sas-3, Rue des Gaudines, 78102 Saint-Germain-en-Laye, France; [email protected].

Pall, Allegro, iCELLis, Mustang, and Seitz are trademarks of Pall Corporation. Accelerator is a service mark of Pall Corporation. Zolgensma is a registered trademark of Novartis AG.