March 25, 2024

A different approach to adeno-associated-virus (AAV) clarification

By Cytiva

Where and when to start clarification process development

Depth filters in combination with membrane filters provide a cost-effective and scalable solution to AAV harvest clarification, particularly when compared to centrifugation, which requires high capital cost upfront and can be challenging to scale-up. The wide range of chemistries and pore sizes of depth filters also mean that a tailored production process can be developed for just about any AAV process.

However, the wide range of options also creates a challenge to find the best combination of depth and membrane/sterile filters. A practical starting point is identifying the optimal combination by understanding the feed stream and process constraints.

Understanding the feed stream

The nature of the feed stream will have a direct impact on the type of depth and sterile filters required for clarification. Factors to consider include the type of cell line, the cell density and viability, whether the viral particles were secreted or released via cell lysis, and the turbidity of the crude harvest. Whether an adherent or suspension process was used is also important.

The crude harvests from suspension cultures generally contain more biomass than those from adherent processes. In these cases, a filter configuration with two or even three layers of filter media can be beneficial; having a dual depth filter with a top more open layer enables clearance of larger debris. This approach is also often beneficial for processes in which cell lysis is not required. The higher load of cells (e.g., large particles) requires an additional open depth filter sheet.

For harvests that contain fine particles, it is important to have a tight depth filter media grade, with a second and possibly third layer added on top to improve filtration capacity and to lower the overall process cost. Conversely, feed streams that are relatively clean (< 10 NTU) may not require a depth filtration step at all.

Process constraint and goal considerations

In addition to the properties of the feed stream, it is essential to consider the process constraints when selecting an initial set of depth/sterile filter combinations. Factors that can affect the clarification process selection include the batch volume, process time or temperature, and the space available for the clarification step.

Target properties for the clarified harvest will also impact the process goals and thus the choice of depth and membrane filters best suited for an AAV clarification process. The most important property is typically the turbidity, which reflects the amount of cellular debris in the clarified harvest. The clarity of the harvest will often impact the performance of the downstream chromatography and other purification steps.

Filterability studies and optimization

With knowledge of the feed stream, any process constraints, and goals, it is possible to select some depth and membrane filter combinations for initial bench-scale filterability studies. At this time, different combinations of depth filters with varying pore sizes and chemistries are evaluated in combination with common membrane filters with respect to capacity, filtrate quality, and product yield.

Once the best filter train option is identified, the process can be optimized by adjusting the process conditions (flux, pre/post recovery flush, etc.), and the scalability can be confirmed.

Case study 1: clarification of an AAV suspension cell culture harvest

We tested our approach to rapidly develop an effective direct-flow filtration process for the clarification of suspension harvest at a customer site. The crude harvests from AAV suspension cell culture processes in HEK293 cells had high turbidities of 430 and 540 nephelometric turbidity units (NTU) after cell lysis.

The goal of the initial bench-scale filterability studies was to achieve an acceptable throughput and viral vector yield for primary clarification. Due to the high initial turbidities, multiple layer depth filter solutions were explored in combination with a Supor™ EAV 0.2 μm membrane filter to reduce potential bioburden.

For the 430 NTU crude harvest, we tested a dual layer filter (the Seitz™ HP PDH11 filter, which is made up of a coarser layer of Seitz K700P media and a finer layer of Seitz V100P media) against a single layer capsule of the Seitz V100P media and another single layer capsule of Seitz Bio 10 media (all three were used in conjunction with the Supor EAV filter as mentioned above). The results were that the dual layered Seitz HP PDH11 filter and the single layered Seitz Bio 10 filter had similar recoveries, but the dual layer had the highest throughput – this was due to the coarser Seitz K700P media layer retaining contaminants in the range of 6 to 15 μm, which protected the finer layer of the filter. However, the Seitz Bio 10 filter did exhibit the highest yield.

In the second filtration study with 540 NTU crude, a Seitz PDP8 dual layer filter (a coarse Seitz T1500P media on top of a finer Seitz K700P media layer) used in series with the Seitz Bio 10 filter followed by the Supor EAV membrane filter afforded the highest throughput and viral vector yield. The Seitz HP PDP8 filter protected the finer Seitz Bio 10 single layer filter and improved the throughput on the Seitz Bio 10 filter by five times that of the Seitz V100P filter alone. Meanwhile, the throughput of the original dual layer Seitz HP PDK11 filter was approximately four times higher than the Seitz V100P filter alone.

The throughput or yield of a dual layer combination needs to be higher than that of a single layer depth filter to make economic sense. While both dual and triple layer options provide viable clarification options (yields of 75–90%), considering the filter area per capsule and the number of filtration steps along with the throughput (230 L/m2) and yield (90%), it was concluded that the filtration process consisting of the dual layer Seitz HP PDP8 filter in series with the Seitz Bio 10 and Supor EAV filters provided the best overall performance. It should be noted that the simpler Seitz HP PDH11 + Supor EAV filters solution did provide good performance for cleaner feeds with lower cell densities.

Case study 2: clarification of an AAV adherent cell culture harvest

In a second case study, clarification of the crude harvest from an AAV5 adherent cell culture process was evaluated. Cell lysis was also performed in this process, leading to the formation of many small particles. In this instance, the crude harvest used for the bench-scale filterability studies had a turbidity of 40 NTU, indicating the need for a combination of depth and membrane filters.

Since the turbidity was not extremely high, a single layer (V100P), relatively tight, depth filter with a pore size of approximately 2–4 μm that was specifically developed for virus filtration was evaluated along with a dual layer (K900P + V100P) option, both of which were used in combination with a commercial, 0.65/0.2 μm membrane filter (Supor EKV).

The dual-layer depth filter included a more open sheet on top of the V100P layer to clear out larger debris and provide for a higher capacity. While both filter trains hit their yield and turbidity targets (>95% and < 5 NTU), capacities of greater than 500 L/m2 and > 2500 L/m2 were observed for the dual depth and membrane filters, respectively – approximately five times that of the single-layer filter without any negative impact on yield, which was 97% recovery across the entire dual-layer filtration train. This solution was therefore determined to be the better option.

This process was scaled for the clarification of crude harvest from a 10 L bioreactor. The 22 cm2 Supracap™ 50 depth filter capsule was scaled to a 254 mm (10 in.) Supracap 100 depth filter capsule and the 3 cm2 syringe membrane filter was scaled to a 220 cm2 Mini Kleenpak™ membrane filter capsule. Despite a much dirtier crude harvest (130 NTU), the performance of the filtration train was largely unaffected with respect to turbidity reduction (< 5 NTU in clarified pool) and yield (95%); only the filter capacity was reduced, as expected.

Importantly, both the footprint and cost for the dual layer solution were determined to be very reasonable. Given the known scalability performance of Cytiva’s bench, intermediate and commercial scale depth filters, it was determined that for a similar crude harvest from an adherent AAV cell culture process performed in a 500 L bioreactor, a single Stax™ depth filter capsule and a single 254 mm (10 in.) Kleenpak Nova sterile filter capsule would be sufficient.

Learn about our AAV solutions.