Workflow for AAV manufacturing
Workflow step What to consider Strategies
Process development Transfection efficiency in HEK293 and SF9 cells; viral productivity; maintain viral infectivity Fine-tune transfection conditions for scalability; optimize cell culture media and expansion; consider developing a stable producer cell line
Scale-up Seamless batch-to-batch scale-up, including transfection; prevent contamination; grow virus for high titers under BSL-2 conditions Digitally enabled bioreactors for data-driven control; single-use solutions to increase efficiency and reduce turnaround time; optimize media with scale
Buffer exchange and concentration
Labor-intensive multistep process with risk of failure; filter fouling, clogging, and contamination risk; product recovery
The right filter media and sizing for scale-up; a scalable tangential flow filtration (TFF) system; single-use, closed systems to reduce steps and handling
Capture
Capacity, recovery, coelution of virus and host cell DNA, and separation of empty/full capsids; affinity purification is challenging due to multiple serotypes
Modern chromatography resins for faster throughput; prepacked formats to control bioburden and reduce changeover between campaigns; standardized column platform from PD to full-scale manufacturing
Polishing
Removal of impurities and separating empty from full capsids; capacity, recovery, and coelution of virus and host cell DNA; complex process, low loading/binding capacity for size exclusion chromatography (SEC)
Optimal balance between capacity and resolution; core beads for efficiency and productivity in size-based separation; flow-through mode where possible to reduce process time
Analytics
Time-consuming assays/imaging; batch-to-batch variation in assay results
Inline assays for real-time measurements; auto sampling/high-throughput systems that can take multiple measurements; multiple assays – qPCR, ELISA, SPR, HPLC, TEM, etc.
Learn life science topics, such as chromatography and bioprocessing, from online learning.
Biotech Makers

In 2020, Cytiva scientists hosted several webinars on developing a scalable adeno-associated virus (AAV) production process, with a focus on AAV serotype 2 (AAV2). They detailed the AAV production workflow, including transient transfection and cultivation in single-use bioreactors, AAV purification, and analytics. And they also previewed a new purification technology. Here they answer questions from attendees.

How long does it take to adapt adherent cells to suspension for the HEK 293T cells?

It took about 2 months using direct adaptation. We made 10 cell passages to control cell growth before creating a working cell bank. One tip is to control cell aggregates during adaptation and not exceed 2.5 × 106 cells/mL, because aggregates will increase when viable cell density (VCD) is too high.

Do you expect your PEI-based transient transfection method for AAV will be scalable to 500 L?

Yes, this is what we expect. But so far for AAV2 we have only done the scaling up from 20 mL to 20 L. We obtained similar titers for several serotypes. A customer has a scalable transfection procedure in the Xcellerex XDR bioreactor up to 500 L.

How did you select your lysis conditions for AAV2?

We have been doing a study comparing different detergents for lysis of HEK293 cells in an adenovirus process that we published and compared with freeze-thawing. Tween™ 20 showed good performance. It is not on the REACH list (i.e., it is health and environmentally friendly and can be used for large-scale production).

Do you have any recommendations to maximize recovery on Capto AVB?

Yes, a few things. Make sure you do not underload the column. Capacity is very high, approx. 1014/mL resin. If the titer is very low in the harvest, try to concentrate it. And try different elution procedures (flows and buffers). Also, the titer in the sample that is loaded can affect recovery (higher titer is an advantage).

The AAV recovery from the concentration and buffer exchange step was up to 80%. Where was the loss of 20%?

We did analyze the permeate, and no virus was detected there for the 300 MWCO hollow fiber filter. It could be trapped to some degree in the system tubing or filter, or it could also be analysis error. Sometimes we have achieved 90% to 100% recovery. So, we generally get very good recovery over this step.

How does Capto Core technology work for AAV purification?

This is a flow-through procedure. The virus will pass through due to the size, and the impurities will go into the pores and bind the ligand (octylamine) inside. We have Capto Core 400 and Capto Core 700. Capto Core 400 is suitable for AAV. The 400 and 700 reflect the MWCO of the outer shell of the bead.

How do people manage viral clearance studies in gene therapy products and processes (e.g., AAV)?

It’s important that the raw materials are not contaminated. But depending on the virus to be purified and what it can withstand, you can also introduce steps with low pH or detergent, so that more sensitive viruses will be cleared.

What is the best way to remove empty AAV capsids?

By ion exchange (IEX) chromatography if you want it to be scalable. We are still working on the full/empty separation, but we believe Capto Q ImpRes is suitable for most serotypes. We also suggest you maximize the % full capsids in your preparation by optimizing upstream production.

What method do you use for measuring AAV titer?

We are using two similar assays: an enzyme-linked immunosorbent assay (ELISA) and a Biacore assay for surface plasmon resonance (SPR). We get very similar values that compare to the ATCC reference standard. But we have lower variation (CV%) with the Biacore assay.

Did you check residual DNA impurities in your AAV preparation?

So far, we have done PicoGreen™ analysis for total DNA. After the affinity step we could not detect any DNA; it was below the detection limit (1 ng/mL). We plan to perform a qPCR for HEK293 DNA that we have already set up. In the AAV DNA qPCR, we always check background from plasmids and have carefully set up the assay to have full control of any background sígnals from plasmids. We see no or very low background in our qPCR assay.


Get details on our AAV process and hear more Q&A ― Sign up here.