Bioprocess Strategies to Efficiently Purify Viral Vectors at Scale

This is an exciting time to work on emerging therapies. Viral vectors are driving the development of many of these new options, mainly cell and gene therapies but also oncolytic therapies and vaccines. Together, these technologies treat rare genetic diseases, cancers, and have the potential to mitigate and control outbreaks and the spread of disease. The promise of gene therapies is relatively untapped, with potential indications expanding to more common diseases. However, because of dual needs to (1) quell lingering concerns over safety, efficacy, purity, and consistency and (2) generate vectors at scale, particularly for gene therapies aimed at larger populations, this article will discuss ongoing efforts to optimize steps after harvest from host cells and to effectively scale up the process for gene therapy vector production. 

AAV-based gene therapy: a problem of scale

The goal of gene therapy is to treat genetic diseases by repairing or replacing deleterious sequences in affected cells with intact copies. The FDA has approved a growing number of gene therapies since 2017, including therapies to reverse a form of inherited blindness and to treat spinal muscular atrophy in children. These approvals illustrate the safety and efficacy of current vector-based approaches. Consequently, many more gene therapies will graduate therapeutic pipelines within the next few years, with up to 20 annual approvals expected by 2025. Viral vectors require high purity and reproducible production from batch to batch. To achieve these requirements amid growing demand, more efficient production of viral vectors at greater volumes is needed.

Non-pathogenic adeno-associated viruses (AAVs) and lentiviral vectors make up the majority of current gene therapy trials and regulatory submissions, with AAVs most common for direct (in vivo) gene delivery into recipients and lentiviruses used for applications where cells are genetically modified outside the body before being administered. AAVs are non-enveloped, single-stranded DNA viruses. With a favorable safety and immunogenicity profile and potential long-term expression, AAVs can be cultivated in culture plates, shaking flasks, or bioreactors using commonly available cell lines without the need for potentially cytotoxic helper virus systems. However, the unique tissue tropisms of different serotypes present both benefits and challenges. 

Legacy downstream purification technologies for making viral vectors in labs combine chemical precipitation and centrifugation based on density gradients. Chromatography has been investigated as an alternative, using various principles of biochemical interaction, including size selection, hydrophobicity, and affinity binding. Purification via centrifugation is manual and not well suited for producing large quantities of virus in a manufacturing environment. Workflows with many or suboptimal chromatography steps can compromise both quality and yield, which is costly and inefficient. When working with AAV, care must be taken to ensure developed protocols work for the AAV serotype(s) of interest. 

It is no easy task to develop downstream processes that check all the boxes for yield, purity, and scalability. Viruses reside among cell lysate debris and other impurities. Therefore, viruses must be singled out from among potentially hundreds of proteins and other biomolecules and then matched to an appropriate purification regime. The final product must be enriched for full, biologically active viral capsids. Both expertise and persistence are required to pick and polish viral targets in this manner. Because of growing pressures to improve purification efficiency while scaling up, a case-by-case approach is rapidly becoming inadequate, and the community is working toward universal protocols. Although these are still being refined, a combination of filtration technology and ion exchange chromatography (IEX) are in use.

A vector purification workflow

Production proceeds via the following steps: First, the molecular machinery of a host cell that contains viral vector genes and the gene of interest repurposes itself towards producing excess quantities of viral particles. During the viral vector purification workflow, the produced particles must be accessed, targeted for capture, and then polished to make a final therapeutic-grade product. Then, the preparation is formulated into the required buffer during a final filtration step. 

When it is time to harvest the virus, host cell lysis allows access to the virus, creating a suspension of debris and impurities that must be reduced to acceptable levels. Mechanical lysis and centrifugation are typically used on the lab bench to lyse cells and clear cellular debris. However, for industrial applications, chemical lysis and DNase digestion followed by depth filtration is preferred for clarification prior to target capture due to the simplicity, reproducibility, and scalability. 

Tangential flow filtration (TFF) is often required to concentrate a clarified lysate without compromising the integrity or infectivity of the viruses. Enveloped viruses such as lentivirus are particularly shear-sensitive so require extra care in handling. Regardless of virus type, thoughtful choice of filter material, pore size, and filtration conditions is required to effectively retain virus particles that have a tendency to squeeze through pores that would seem to be the appropriate size. Hollow fiber filters are a common choice for this step as they are gentle enough for labile viruses and are available in a wide range of sizes for ultimate scalability. Automated filtration systems minimize manual handling, allow precise control of conditions, and support reproducible processing. 

Capture and polishing come next in the viral vector purification process. The downstream process must clear contaminants such as host cell protein and DNA to a level that aligns with regulatory requirements and must enrich full capsids, which contain the complete genetic complement, to a high level. Viral particles may be captured using an affinity resin or other interaction. A common approach is to alternate different types of ion exchange. This is particularly true for viral vectors that do not have an affinity option available. A popular choice for AAV polishing uses a strong anion exchange (AIEX) matrix with a quaternary amine as ligand. This method exploits the slight difference in pI of the desired full capsids and empty capsids, with pI of 5.9 and 6.3, respectively. When developing an AAV process, it is critical to ensure that it is suitable for the various serotypes of interest, with some fine-tuning expected especially with the polishing step. For lentiviruses, a weak anion exchanger for polishing may enhance recovery.

In summary, much progress has been made towards a more standardized platform for viral vector purification that will meet the needs for large-scale industrial production of the future, though refinements are still being explored especially for AAV.