Cost-effective manufacturing of r-AAV using Pall’s Allegro™ STR bioreactor
Poster Content as Presented at BPI Europe & ESACT Conferences 2022
Marie-Laure Collignon1, David Mainwaring1, Andreia Pedregal1, Adrien Auffret-Cariou2, Brian Mullan2, Cédrick Rousseaux2 & Sophie Blondel2
1 Pall Corporation, Reugelstraat 2, 3320 Hoegaarden, Belgium; 2 Yposkesi, an SK pharmteco company, Rue Henri Auguste-Desbrueres, 91100 Corbeil-Essonnes, France
To date there is huge pressure on biomanufacturing facilities due to the high demand for the manufacturing of gene therapy products such as recombinant adeno-associated viral (rAAV) vectors under current good manufacturing conditions.
Yposkesi, part of SK pharmteco, is a leading contract development and manufacturing organization in Europe for rAAV and LVV production and is investing significantly to improve the productivity of viral vector manufacturing processes.
Yposkesi has evaluated Pall’s Allegro STR 50 bioreactor for its usability and performance by conducting a proof-of-concept study with a triple transient transfection rAAV process, using a proprietary transfection reagent, a high-productivity HEK293T cell line, and a custom medium.
The process tech transfer was performed by keeping the power input per unit volume (P/V) and the superficial gas velocity constant. The initial process conditions resulted in foam production and required the addition of some antifoam. Slower cell growth and a lower AAV yield was obtained, compared to the original STR platform.
The cell culture parameters were then re-developed by keeping a constant low sparger air flow rate and only oxygen in the dissolved oxygen (DO) cascade, to avoid foam generation. The resulting yields were significantly higher than the original STR platform.
Yposkesi compared and experimented with different suppliers of bioreactor.
The user-friendliness of Pall’s Allegro STR bioreactor is a strong point, for both biocontainer installation and software operations, and the close collaboration between Yposkesi and Pall scientific supporting teams resulted in a favorable increase in production yields.
Moving forward, the process is planned to be scaled up to 1,000 L, in order to further increase cost-effectiveness of large-scale manufacturing for rAAV vectors.
- Cell line: Yposkesi proprietary HEK293T cell line
- Cell culture media: Custom F17 media
- Transfection reagents: A mix of AAV plasmids (pRepCap, pHelper, pGOI) and a proprietary polymer for plasmid complexation diluted in cell culture media
- Nuclease reagents: A mix of Benzonase♦ and MgCl2 diluted in cell culture media
- Lysis reagent: Chemical lysis agent
- Bioreactor: Allegro STR 50 bioreactor (Pall) with Lauda♦ Variocool♦ VC 200 thermal control unit (Lauda-Brinkmann)
- Analytics: Vi-CELL♦ cell counter (Beckman Coulter), BioProfile♦ FLEX2 cell culture analyzer (Nova), AAV vector genome titration (qPCR)
- HEK293T cells were thawed and expanded in shake flasks for 11 days
- Allegro STR 50 bioreactor was inoculated with 0.25 x 106 cells/mL, with a three day cell expansion phase in the bioreactor
- Triple plasmid transfection and DO set point modification (50 to 10%) were carried out at Day 3, DNase treatment at one day posttransfection (Day 4) and chemical lysis at three days post-transfection in the bioreactor (Day 6)
- 50 L of AAV bulk was harvested after cell lysis
|Tech Transfer Parameters||Competitor STR 50 Bioreactor (Scale Down STR 200)||Initial Allegro STR 50 Bioreactor||Optimized Allegro STR 50 Bioreactor|
|Production working volume||50 L||50 L||50 L|
|Power input per volume||Similar (correlated agitation speed)||Similar (correlated agitation speed)|
|Direction of agitation||Downflow||Upflow||Downflow|
|Sparger air flow rate||Air injection increasing (DO output 0 – 60%)||Low constant air flow|
|Sparger O2 flow rate||O2 injection increasing (DO output 60 – 100%)||Gradually O2 injection increasing (DO output 0 – 100%)|
|Total sparger flow||Similar superficial gas velocity||4–5 times less gas hold up|
|Anti-foam addition||Not needed||Required||Not needed|
The agitation condition in the initial process and optimized process in the Allegro STR 50 bioreactor was defined by keeping the same power input per unit volume (P/V) as the competitor STR bioreactor.
Also, the sparger aeration conditions of the initial process in the Allegro STR 50 bioreactor kept air and O2 in the DO cascade loop the same as the competitor STR bioreactor, and converted air and O2 sparger gas flow range to maintain similar superficial gas velocity and similar kLa. This led to unsatisfying results.
Going forward, the optimized process in the Allegro STR 50 bioreactor was redesigned by applying a constant low sparger air flow rate and keeping only oxygen in the DO cascade loop.
- AAV viral genome copies were measured in cell lysed bulk samples.
- The initial process in the Allegro STR 50 bioreactor mimicked the Yposkesi process as closely as possible but led to a decrease of about 75% in AAV titers.
- The optimized process in the Allegro STR 50 bioreactor was successful as the AAV titers achieved was twice the level achieved with the competitor STR bioreactor.
O2 and Air Sparge
During the initial process in the Allegro STR 50 bioreactor, an excessive foam generation was observed. It was a consequence of getting sparger air and O2 flow rate in the DO control loop and converting flow range at same superficial gas velocity as the competitor STR bioreactor, to target same kL a. Indeed, this methodology translated in increased about 40% vvm values compared to the competitor STR (data not shown).
The optimized process in the Allegro STR 50 bioreactor kept only oxygen in the DO cascade loop and applied constant low sparger air flow rate. This led to an efficient process with no foam generation and no need to anti-foam.
Indeed, as shown on the graphic, the amount of total gas added (air and O2) in the initial process was 4 to 5 times the amount in the optimized process, which might have induced shear stress and has required anti-foam addition. Those could potentially explain the low AAV titers initially obtained.
Cell Growth and Viability
During the growth phase (Day 0 – Day 3), comparable and consistent cell growth and viability profiles were observed for the optimized process in the Allegro STR 50 bioreactor runs and the process in the competitor STR50 bioreactor with a viability above 95%. After transfection, cell growth rate was slower in the Allegro STR 50 bioreactor compared to the competitor STR bioreactor. Cell viability drop was a similar way.
Cell growth and viability in Allegro STR 50 bioreactor (optimized process) and competitor STR50 bioreactor
During growth phase, glucose and lactate profiles were comparable and consistent between the optimized Allegro STR 50 bioreactor runs and competitor STR50 bioreactor.
After transfection, the consumption rate of glucose and production rate of lactate in the Allegro STR bioreactor was slower compared to the competitor STR. This correlated with the cell density profile.
Glucose and lactate levels in Allegro STR 50 bioreactor (optimized tech transfer) and competitor STR50 bioreactor
pH and Dissolved Oxygen
The pH process value was maintained at 7.2 during the first two days of culture, then dropped slightly to values of around 7.0 – 7.1. This was a consequence of medium acidification resulting from cellular metabolism and production of byproducts as lactate.
In the optimized process in the Allegro STR 50 bioreactor, the pH did not increase after 90 hours, probably due to the low constant air flow.
Additionally, a very comparable and consistent DO regulation was registered for Allegro STR 50 bioreactor runs and competitor STR50 bioreactor.
pH and DO in Allegro STR 50 bioreactor (optimized tech transfer) and competitor STR50 bioreactor
The proof-of-concept of Yposkesi rAAV process was successfully transferred from a competitor STR to Pall’s Allegro STR 50 bioreactor platform with an optimized 2-fold product yield.
Maintaining a constant P/V, constant low sparger air flow rate and only oxygen in the DO cascade loop led to an efficient process with no foam generation and potentially less shear. This could have led to a change in cellular metabolism and explain the titer increase.
In addition, usability of the Allegro STR 50 bioreactor was also evaluated by Yposkesi. Compared to other suppliers tested by Yposkesi, the ease of the biocontainer installation and user-friendliness of the software were highlighted.
Moving forward, the process is planned to be scaled up to 1,000 L in order to further increase cost-effectiveness of large-scale manufacturing for rAAV vectors.
This case study highlights that rAAV process can be moved between different bioreactor supplier platforms, while increasing process performance. This could have benefits for product developers who may feel constrained to a given bioreactor supplier.