From Bench to Bedside: A Scalable End-to-End Solution for AAV Production
Posted on June 20, 2022
Peiqing Zhang, PhD, Strategic Technology Partnership Leader APAC, Pall Corporation
Nicholas Marchand, PhD, R&D Manager, Pall Corporation
Mark Schofield, PhD, Sr. R & D Manager, Pall Corporation
Debbie King, Scientific Technical Writer, Cell Culture Dish
Brandy Sargent, Editor in Chief, Cell Culture Dish
The need for industrialized, cost-effective virus manufacturing platforms is growing rapidly as their usage in precision medicines continues to rise. Among viral delivery systems, recombinantly produced adeno-associated viruses (AAV) are now the predominant vector for in vivo gene therapies with good safety and gene transfer profiles. These viral vectors transport genetic information to a target cell, thereby delivering functional genes needed to correct a genetic defect, to inhibit the cell from producing a defective version of the gene and restore normal function or to impart new function to the cell to treat (cancer treatment) or prevent disease (vaccine). One of the biggest challenges in bringing these life-changing therapies to patients is the ability to efficiently translate bench-scale processes to commercial GMP manufacturing to produce viral vectors at the scale needed to meet dosage requirements. Specifically, as more prevalent indications are being tackled, such as Duchene’s Muscular Dystrophy (DMD), where high vector doses are required (>1 x 1014 vg/kg), manufacturing is a significant bottleneck1. Such high demands for viral vectors highlight the need to implement easily scalable manufacturing processes while also increasing overall productivity and managing cost of goods. Manufacturers are tasked with developing technologies to produce vectors at the right quantity, quality, and cost to support pre-clinical to clinical development, and ultimately, commercial stage manufacture.
Many of the manufacturing technologies used in early-phase viral vector development are bench-scale processes. While these deliver high efficacy at small scales, they are neither robust, nor effective, when scaling towards large-scale GMP manufacturing. New technology solutions to meet comprehensive program requirements from early-stage to commercialization have been recognized by suppliers, who are bringing fit-for-purpose scalable platforms tailored for viral vector production. Here, the benefits of adopting an end-to-end manufacturing solution can be observed. Instead of patching together disparate equipment and technologies from different vendors, utilizing the expertise and technological capabilities from a single end-to-end solutions provider, such as Pall, is an efficient way to implement these strategies into existing workflows and support unmet process needs. Creating comprehensive process solutions requires input from multi-disciplinary teams including skilled process, automation, and engineering experts, and needs to be supported by a wide-ranging product portfolio and performance data. Pall utilizes this holistic approach to create a single harmonized manufacturing solution aligned with good manufacturing practice (GMP) for AAV production based on process needs, budget and facility/cleanroom space (Figure 1). Manufacturing solutions with single-use technologies that can span all the unit operations from upstream through downstream also ensures continuity and compatibility of equipment and consumables while simplifying validation and qualification processes. Integrated systems can support rapid scale-up, streamline operations end to end, speed time to market, improve process economies, and maximize the chance for overall project success by controlling risk through proven platforms and designs.
Any end-to-end manufacturing solution needs to be under-pinned by an intricate understanding of the process needs. Although the precise process requirements will vary for each individual therapy, the core process platform can be adapted and tailored based on proven workflows.
Here, we will look at a typical AAV production process utilizing a feasibility study conducted to evaluate the performance of a platform process for an AAV5 viral vector production and purification. Enabling technologies from Pall were used for each stage of the process to demonstrate a contiguous manufacturing platform with scalable performance for AAV. This platformable approach was modeled after traditional monoclonal antibody workflows where the same platform can work for multiple mAbs. (For more detailed information on the study design and experimental details, please contact Pall).
Viral Vector Production
Transient transfection of mammalian cells like HEK293 with multiple plasmids was the first and remains the most widely used method to produce AAVs2. Transient transfection is attractive for its flexibility and speed, thus advantageous for early stages of investigational product development. However, there are challenges with this method, such as low vector yield, that need to be overcome to support successful transition to late-stage production. Overcoming transfection challenges is an active area of development and great strides are being made where transfection reagents optimized for AAV vector production could provide significantly higher yields than current methods3.
Adherent vs. Suspension
Adherent HEK293 cells have been widely adopted for clinical manufacturing of AAV vectors but there is a fundamental lack of scalability due to the large number of flasks, roller bottles, or multi trays—and process control—required for a clinical production run4. One solution is to adapt these cell lines to grow in suspension and to utilize existing intensification strategies for established biologics, including stirred-tank bioreactors (STR) for large volume production. Stirred tank bioreactors are offered in a variety of size and configuration options. For example, the Allegro™ STR bioreactor (Pall Corporation) is a single-use bioreactor that is available in 50, 200, 500, 1000 and 2000 L scales. The wide range of models offer flexibility of choice in production scale as well as a convenient path from cell seed generation through seed train. A critical consideration in selecting bioreactors for up-scaling of production in addition to performance is scalability through the sizes. Studies have shown linear scalability from Allegro STR bioreactors from 50 L to 500 L for the production of AAV from HEK293 suspension cells and that there is no significant difference between the two bioreactors as long as the same process is performed. In a recent study for process scalability, 50 L and 500 L working volumes were evaluated based on key process parameters such as cell growth, viability, metabolic profile, and vector production5. Data demonstrated the scalability of a transfection-based AAV production process between the Allegro STR 50 and STR 500 bioreactors. Analysis of key process parameters showed similar growth kinetics, viability, and vector titers throughout the evaluation period. While slightly higher titers were observed in the Allegro STR 50 bioreactor, the productivity between the two scales was within ~10% indicating a scalable process.
Ultimately, not all host cell lines can be successfully adapted to grow in suspension culture while maintaining high productivity and quality. Additionally, the extra time for cell line development and cost may not fit within the condensed developmental timelines imposed on high demand AAV products. Advances in bioreactor technology are supporting adherent processes. For instance, the iCELLis® fixed-bed bioreactor system from Pall, with linearly scalable formats, can be effectively used to span requirements from small-scale early-stage (iCELLis Nano bioreactor, 0.53 to 4 m² surface area) to late-stage clinical and commercial production volumes (iCELLis 500+ bioreactor, 66 to 500 m² surface area). The iCELLis family offers enhanced volumetric productivity, along with integrated process controls (i.e., pH, dissolved oxygen (DO), and biomass) for easy process transfer and scalability from bench-top to industrialized production of AAV. For common host cell lines, iCELLis 500+ bioreactor at the 500 m2 scale demonstrates comparable vector yields to a 500 L – 2000 L suspension-based stirred tank reactor. Enabling technologies such as this have gained increasing acceptance in the biopharmaceutical industry. For example, iCELLis fixed-bed bioreactors are used to produce FDA approved Zolgensma® (onasemnogene abeparvovec-xioi), an AAV-based gene therapy for spinal muscular atrophy.
FEASIBILITY STUDY RESULTS
Adherent HEK293T cells from a qualified master cell bank were used to establish a seed train in T-flask and multi-layer trays prior to biomass transfer to the iCELLis Nano bioreactor. Cultures were maintained until the cell density reached ~150,000 cells/cm2 at which point transient transfection was performed. Subsequently, cell lysis was performed five days post-transfection followed by downstream purification of the crude harvest with an average titer of 7.82 x 109 gc/mL.
Clarification is a critical step necessary to remove the host-cell impurities (including cell fragments and some proteins and DNA) from the culture lysates to reduce the biological burden and improve viral yield downstream. The easiest and most economical technology to clarify the cell culture is filtration. The appropriate filter or combination of filters for large-scale production needs to combine high capacity for impurity removal, high product yield, ease of scale-up and good overall process economics to find the “best fit” strategy for the specific process.
In one study, multiple combinations of depth and sterilizing-grade filters were screened to evaluate clarification performance6. The screening results showed the combination of Pall’s P-grade PDK11 depth filter in series with a Supor® EKV sterilizing-grade filter resulted in the highest performance based on capacity, impurity reduction, and yield. This PDK11/EKV filter train strategy was utilized on AAV5 adherent cell culture harvest with an initial turbidity of 40 NTU and effectively reduced turbidity to < 5 NTU at 5X higher capacity compared to single-layer filter train without any negative impact on yield (97% recovery). Given the known scalability performance of Pall’s bench, intermediate and commercial scale depth filters, it was determined that scaling up to a similar crude harvest from a 500 L pool coming from an iCELLis 500+ bioreactor could theoretically be harvested using a single Stax™ depth filter capsule and a single 254 mm (10 in.) Kleenpak™ Nova sterile filter capsule. This highlights how the direct filtration approach can be an effective, scalable, and cost-effective clarification solution for large-scale AAV vector production6.
FEASIBILITY STUDY RESULTS
The PDK11/EKV filter train was evaluated for robustness over the course of eight bioreactor harvests. The filter train was able to consistently reduce the crude harvest turbidity. Strong process robustness against feed stream turbidity with regards to clarified pool turbidity and step yield was observed. Filter capacity was influenced by feed turbidity; however, all eight runs showed >250 L/m2 and >450 L/m2 throughput on the depth filter and sterile filter, respectively.
In downstream processing, concentrating the feed after clarification can be implemented to shorten the loading time required for affinity chromatography purification.
FEASIBILITY STUDY RESULTS
Pall’s Cadence® single-use modules with Omega™ 100 kDa PES membrane were evaluated across six samples to concentrate the clarified AAV5 pool to a target volumetric concentration factor (VCF) of 10X. Flux excursion testing reported critical TMP at ~10–15 psi, AAV retention at ≥99.7% across the range of operating conditions, with no significant virus loss due to shear at crossflow rates up to 7.5 L/m2/min. An average vector step yield of 91% was observed with a 95% confidence interval of ± 8.0%.
Traditionally, viral vectors have been purified using ultracentrifugation for research or small clinical studies, but there has been a move toward chromatography, considered to be a more scalable and cost-effective strategy to support scale up into the clinic. Affinity chromatography has become the method of choice for AAV selective capture, and different affinity ligands are available for specific capture of various AAV serotypes. For AAV, AVB Sepharose™ affinity resin (Cytiva) utilizes a recombinant protein ligand and has been shown to be an effective strategy for the purification of several AAV serotypes7. The initial chromatography serves as an initial bulk cleanup step after clarification and is typically followed by an anion exchange chromatography (AEX) polishing step since the affinity purified feed still contains residual impurities.
FEASIBILITY STUDY RESULTS
Across 16 total affinity chromatography purifications AAV5 recovery was determined to be 68 ± 13% by capsid ELISA and 57 ± 30% by ddPCR method.
Polishing by Membrane Chromatography
The biggest challenge during AEX purification is separating capsids that contain the gene of interest from closely related impurities including empty capsids. Debate remains as to the impact of these empty capsids, but they are regarded by regulatory authorities as in-process contaminants and downstream removal is a necessary step to achieve a consistent drug product. At laboratory scale, analytical ultracentrifugation (AUC) is effective to purify full AAV capsids, but this technique is not practically scalable. Recent advancements are improving the ability to enrich for full capsids, such as Pall’s new patent-pending elution method using the Mustang® Q membrane AEX chromatography8. Instead of classical linear gradients, small conductivity-steps of ~1 mS/ cm yields distinct elution peaks for both “empty” and ”full” capsids, offering rapid visual readout to the success of the purification and rapid path to process development. Method testing demonstrates efficacy at small scale using 0.86 mL Mustang Q XT membrane in the Acrodisc® capsule format and effective scalability to the 5 mL Mustang Q XT capsule on the Cytiva ÄKTA™ avant chromatography system. Using the Mustang Q membrane with step gradient elution, it was possible to achieve 4 to 5-fold full capsid enrichment and obtain elution fractions containing ~100% full capsids. By adjusting the conductivity step gradient and elution salt, protocol optimization is possible for different AAV serotypes demonstrating its potential as a truly scalable platform technology well-suited to meet transitioning needs from early-stage to commercial manufacturing.
FEASIBILITY STUDY RESULTS
The affinity purified vector pool was polished to enrich for full capsids using Mustang Q XT anion exchange membrane sorbent. Elution fractions were analyzed: Full capsids containing the target gene using ddPCR and total capsids (AAV5 ELISA). A representative chromatogram is shown above (Figure 2) where elution fractions 3, 4 and 5 are enriched for full capsids. To show reproducibility this purification was performed for 5 independent feedstreams, on average the empty peaks (1+2) have a 49 ± 10% capsid yield while the genome (vg) yield is only 14 ± 8% while capsid yield in the full peaks (3+4+5) is 19 ± 10% with genome (vg) yield at 66 ± 13%. When mass balance is calculated from fraction pooling, a close to a 5-fold enrichment of full capsids to total capsids is observed.
Next steps in the viral vector manufacturing workflow involves adjusting the buffer and virus concentration to the final formulation condition ready for clinical use (ultrafiltration/diafiltration (UF/DF)). Tangential flow filtration (TFF) is a scalable approach that is typically employed at this step. TFF solutions are commercially available in flat-sheet cassettes and hollow fiber modules, and both have utility for viral vector manufacturing. An advantage of flat-sheet cassettes like Pall’s T-Series cassettes with Omega polyethersulfone (PES) membranes is they often provide higher flux than hollow fiber modules. Reusable flat-sheet TFF cassettes also offer a flexible ‘modular’ approach to increasing filtration capacity by simply ‘stacking’ more cassettes together in the filter holder8.
FEASIBILITY STUDY RESULTS
Cadence single-use modules with Omega 100 kDa membrane were evaluated with four samples for ultrafiltration/diafiltration (UF/DF) of the purified AAV5 pool with a target of 10X VCF concentration and 7X diafiltration volume buffer exchange into formulation buffer. The AAV5 concentration/diafiltration was performed at a crossflow rate of 7.5 L/m2/min and a TMP of 15 psi. Permeate flux remained steady at ~200 LMH throughout the concentration and diafiltration. Virus concentration measurements in the final permeate pools demonstrate virus retention ≥99.9% and final virus yields averaged 89% over the four trials following UF/DF.
Final sterile filtration is an essential process step to remove bioburden and mitigate the risk of microbial contamination of the final viral vector product to ensure patient safety. In most cases, sterile filtration utilizes 0.2 µm sterilizing grade filters. Pall’s validated Supor® EKV sterilizing-grade membrane filters are often used because of their low protein binding, which ensure the maximum transmission of the active ingredients. As well, the EKV filters are available in a variety of formats to suit a range of processing volumes from the Mini Kleenpak 20 (50 mL – 2 L) to the Kleenpak Nova (100 L – 1000 L) capsule formats, with scalable and robust performance.
FEASIBILITY STUDY RESULTS
A Supor EKV 0.2 µm sterilizing-grade filter in a Mini Kleenpak Syringe capsule (2.8 cm2) in constant flow mode with a flux target of 500 LMH was used for final sterile filtration. Virus concentration was measured in the feed and filtrate pools to calculate transmission.
High virus transmission (averaging 94%), was achieved at final filtration with a Supor EKV filter.
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 effectively scaled up from research scale to meet the quantities required for late-stage clinical and commercial use. For efficient transition between early-stage research and commercialization, it is critical that decisions and investments be made early in the development of manufacturing processes and that the scalability and flexibility of a technology be taken into consideration. These factors are key to meeting evolving process requirements and regulatory demands while being adaptable to changes in the market and emergence of new technologies. Industry-wide initiatives between researchers, drug developers, and solution providers are overcoming limitations in AAV production technologies with fit-for-purpose platform designs to decrease processing time, lower cost, and speed time-to-market. Throughout this article, we have described the key process steps and highlighted the value of end-to-end process solutions. Beyond enabling technologies, Pall’s Acceleratorsm process development services can help guide developers through each step in the optimization of their specific gene-therapy processes with a comprehensive ecosystem of services, novel platform technologies, and proven methods. Coupled together, these technologies and service solutions not only enable production of current therapies but also serve as a valuable blueprint to enable the success of future gene therapy modalities.
FEASIBILITY STUDY RESULTS
The overall theoretical yield from this process is 25% full capsids with a full capsid enrichment of ~5-fold compared to the total viral particles resulting in a purified AAV5 product with host cell protein and DNA impurities below detection levels (Figure 4).
The study demonstrates effective translation of proven, scalable purification technologies used in recombinant protein manufacturing to the purification of viral vectors. An effective end-to-end platform production and purification strategy with technologies from a single provider has the potential to streamline the transition from research to industrialized AAV production.
1. Ayuso E. Manufacturing of recombinant adeno-associated viral vectors: new technologies are welcome. Mol Ther Methods Clin Dev. 2016;3:15049. Published 2016 Jan 6. doi:10.1038/mtm.2015.49
2. Clément N, Grieger JC. Manufacturing of recombinant adeno-as-sociated viral vectors for clinical trials. Mol Ther Methods Clin Dev. 2016;3:16002. Published 2016 Mar 16. doi:10.1038/mtm.2016.2
3. Cameau E. Overcoming Obstacles in AAV Viral Vector Manufacturing. Bioprocess International. June 18, 2021. Accessed April 8, 2022. https://bioprocessintl.com/sponsored-content/overcoming-obsta-cles-in-aav-vector-manufacturing/
4. Smith J, Grieger J, Samulski RJ. Overcoming Bottlenecks in AAV Manufacturing for Gene Therapy. Cell Gene Therapy Insights 2018; 4(8), 815-827. doi:10.18609/cgti.2018.083
5. Sanderson T et al. Scalability comparison between 50 and 500 liter stirred tank bioreactor for production of AAV viral vector Cell & Gene Therapy Insights 2021; 7(9), 1025–1033. DOI: 10.18609/cgti.2021.131
6. Liu M. A Different Approach To Adeno-Associated Virus (AAV) Clarification. May 5, 2021. Accessed April 8, 2022. https://www.pall.com/en/biotech/blog/different-approach-aav-virus-clarification.html
7. Petit S, Glover C, Hitchcock T, Legmann R, Schofield M. Downstream Manufacturing of Gene Therapy Vectors. Cell Culture Dish. January 27, 2021. Accessed April 8, 2022. https://downstreamcolumn.com/down-stream-manufacturing-gene-therapy-vectors-2/
8. Pall Corporation. Full Adeno Associated Virus (AAV) Capsid Enrichment Using Mustang® Q Membrane. Application Note USD3545a/22-07969 01/22. https://www.pall.com/content/dam/pall/biopharm/lit-li-brary/non-gated/application-notes/aav-capsid-enrichment-using-mus-tang-q-an-en.pdf
Copyright 2022 Cell Culture Dish. All Rights Reserved. www.CellCultureDish.com
Register for Webinar on AAV Manufacturing and Scale up here
Subscribe to our FREE newsletter and WEBINAR UPDATES