Gene therapies that are based on adeno-associated virus (AAV) vectors have demonstrated immense potential to treat and even cure previously intractable diseases, but unfortunately the costs are extremely high. AAV gene therapies with high per-patient costs include Hemgenix, a hemophilia B treatment from CSL Behring ($3.5 M); Elevidys, a Duchenne muscular dystrophy treatment from Sarepta Therapeutics ($3.2M); Roctavian, a severe hemophilia A treatment from BioMarin Pharmaceutical ($2.9M); and Zolgensma, a spinal muscular atrophy treatment from Novartis ($2.1M). Luxturna, a retinal dystrophy treatment from Spark Therapeutics, is $850,000 per patient, or $425,000 per eye.
Although these therapies are costly on a per-patient basis, they are not ruinously expensive when spread across the entire population. After all, these therapies focus on rare diseases. But AAV gene therapies of comparable expense for common diseases would be difficult to sustain for even the most generously funded healthcare systems.
If AAV gene therapies are to become more widely available, opportunities to reduce costs must be seized. And the biggest opportunities are in manufacturing. For example, AAV gene therapy manufacturing could become much less costly if it became more like monoclonal antibody and recombinant protein manufacturing through the establishment of platform processes.
What are platform processes? A good definition has been offered by a team of researchers based at the National Institute of Standards and Technology. In a recent article (Biotechnol. Prog. 2023; 39(6): e3378), they wrote that platform processes can be described as “a collection of distinct parts, components, or modules that are recurrent in the manufacturing of a set of products with common characteristics.” They added that “recurrent units permit efficient leveraging of prior knowledge when optimizing the production of a new product with those same characteristics.”
Besides facilitating production development, the platform process approach represents progress toward continuous manufacturing. However, in AAV gene therapy manufacturing, the approach can be hard to follow. Technological obstacles abound. Fortunately, as this GEN article shows, many of these obstacles are becoming less formidable through the development of AAV-specific production technologies.
Process intensification and quality assurance
Demand for large-scale AAV production is increasing due to a high vector dose requirement for efficacy in specific disease states and the expansion of AAV gene therapies beyond rare diseases. Enhanced transfection reagents for mammalian cells in suspension or fixed-bed bioreactors allow large-scale manufacturing processes that deliver high viral titers at harvest. However, the availability of the starting materials used to produce AAV vectors at these scales may limit robust production campaigns.
“Safety of AAVs is a top priority,” says Mathieu Boxus, PhD, global head of bioprocess science, viral vectors, Thermo Fisher Scientific. Careful monitoring is required throughout the manufacturing process to ensure that the capsids have complete genetic material.
Standardization allows for significant production improvements; however, product-specific considerations are necessary for efficacy and safety. Technology that allows for real-time data monitoring and analytics is critical as AAV production processes remain difficult to control and characterize.
“In-depth characterization of the quality along with contaminant analysis requires advanced technologies such as next-generation sequencing and mass spectrometry,” Boxus points out. “Bioassays that are relevant for the intended biological effect beyond generic infectious titer or expression assays can be challenging to develop.”
Although approaches such as process intensification and the use of stable cell lines show promise for addressing increasing demand, they also present challenges. Processes relying on high cell densities usually show limited transfection efficiency, require a large number of plasmids, and can be more costly. However, in the establishment of stable cell lines, the main difficulty is overcoming the cellular toxicity that is due to the expression of vector components.
Manufacturing platforms must address the processing specificities posed by different AAVs—specificities that reflect not only capsid proteins but also genetic payloads. For example, they should facilitate the design of flexible and accelerated processes for specific needs.
To attend to such details, Thermo Fisher relies on validated scale-down models spanning multiple production technologies along with analytical platforms. Doing so facilitates the assessment of key process parameters and supports the rapid scale-up and transfer of processes to cGMP manufacturing.
Sourcing, capacity, and quality considerations
Difficulties in AAV manufacturing include the sourcing of critical raw materials, particularly plasmids (such as pHelper, pGOI, and pRC). “Time- and cost-effective sourcing of all raw materials affects overall program budgets and manufacturing timelines,” states James Cody, PhD, associate director, technical sales and evaluations, Gene Therapy CDMO Services, Charles River Laboratories.
Another difficulty is manufacturing capacity. As the industry matures with additional product approvals, larger amounts of the drug products and raw materials will be needed for commercial supply.
Yet another difficulty is meeting critical quality attributes and achieving lot-to-lot consistency. For instance, a bottleneck can occur at batch release due to the number of quality control tests required to demonstrate product quality and safety.
All young fields need to adopt standardized materials and manufacturing processes. “There are several strategies for AAV manufacturing, both when it comes to upstream/downstream processing and analytics,” Cody points out. “We have invested efforts and resources into the standardization of the process to improve predictability and potentially reduce overall timelines.”
To that end, Charles River recently launched several manufacturing platforms, including nAAVigation for AAV and eXpDNA for plasmids, that offer industry-leading timelines while maintaining product quality. In addition, a set of off-the-shelf AAV-packaging plasmids, including pHelper and multiple rep-cap plasmids, are designed to streamline AAV-based gene therapy programs.
One of the goals of the nAAVigation platform was to help AAV manufacturers meet their serotype requirements. Although there are process differences during downstream manufacturing, the upstream is essentially the same across serotypes and the analytics are standardized. The goal is to make manufacturing as agnostic as possible with respect to serotypes and genes of interest.
Developing optimal parameter ranges
Takara Bio offers an in-house, complete start-to-finish workflow for AAV production that includes process development, plasmid and AAV GMP manufacturing, quality control, and release testing.
“Process parameters between shaking flasks and bioreactors as well as between small and large (>50 L) bioreactors are very different, especially if these devices are from different suppliers” says Michael Haugwitz, PhD, director, cell biology R&D, Takara Bio USA. “Developing an optimal range of parameters for each process is crucial.” Ideally, the scaling up of the bioreactor process should involve minimal optimization.
Interestingly, costs can be high when using a complete system—producer cells, transfection reagents, media, and extraction system—from the same provider. Haugwitz remarks that if you want to reduce costs by using an alternative component, you will need to optimize the complete workflow to accommodate new component. Any workflow change can affect production, purity, and/or the full-empty ratio of an AAV preparation.
Moreover, during the chromatography steps in purification, different serotypes have different binding and elution profiles and require refinements. Yields can also differ depending on serotypes.
Improvements are needed in several areas. Cell lines are needed that can offer higher yields and better full-empty ratios. Transfection reagents are needed that are less expensive and more efficient. Analytical methods are needed that can monitor AAV productivity and control culture medium components and metabolites in the bioreactor. And purification methods are needed that can deliver consistent AAV yields and full-particle ratios over 70%.
Ideally, as Haugwitz points out, manufacturers would be able to rely on serotype-agnostic chromatography methods. In reality, however, optimization is probably necessary, as well as a system that would allow for a serotype-agnostic plasmid ratio for the initial transfection of producer cells and shorter optimization timelines for factors such as gene of interest and serotype.
“Using long-read sequencing could allow for more information about the genome integrity of an AAV preparation,” Haugwitz adds. “Although it is not yet an established method with a defined requirement, it might be of interest to customers as well as regulatory agencies.”
Quantity, quality, and consistency
“We need flexible production systems that fit the manufacturing scale required for the particular disease being targeted,” says Clive Glover, PhD, vice president, viral vectors, Cytiva. For example, the manufacturing requirements for a locally administered treatment differ greatly from those for a systemically administered treatment.
Existing tools can be leveraged to control several impurities that look very similar to those for monoclonal antibodies and recombinant proteins. However, the field is struggling with unique impurities such as empty capsids and encapsidated host cell DNA.
“We need to be able to make the product the same every time,” Glover emphasizes. Consistency is hindered by inherently inconsistent manufacturing processes such as transfection along with variable analytical techniques that make it difficult to gauge consistency. According to Glover, regulatory authorities are currently taking an interest in consistency issues.
The use of stable cell lines, like those used ubiquitously for monoclonal antibodies, may play a pivotal part in resolving quantity and consistency and likely help with quantity. With experience scaling both adherent and suspension processes up to 2,000 L, Cytiva offers cell line services to make bespoke stable cell lines and packaging cell lines as well as process development services.
“It is always best to ensure that the product coming out of upstream is of the highest quality and at the right quantity,” Glover advises. “This will never be completely perfect, so downstream processes also need innovation. If we manage to make significantly more productive systems, then the current downstream workflow will lack the capacity to purify products.”
Each serotype and gene of interest brings different physicochemical properties that will have to be accommodated by changes in upstream and downstream processes. According to Glover, this kind of variation is similar to that seen in the processing of other large molecules such as monoclonal antibodies and recombinant proteins. However, he adds that the variation between products is larger for viral vectors.” He concludes, “All require adaptation of a platform process to a specific molecule.”
Enabling basic research
AAV therapeutics would not exist without basic research. “Our AAV tools enable early-stage research,” says Xuan Liu, PhD, senior vice president and general manager, OriGene. “The main differences between research- and GMP-grade AAV products are the quality and quantity requirements.”
AAV applications of all kinds are limited by the AAV particle’s payload capacity, which is 4.5–4.9 Kb. For example, in CRISPR research, the AAV particle has difficulty packaging CRISPR-Cas9 mRNA, which is about 4.2 Kb. Typically, another AAV particle is needed to package the guide RNA that must complex with the Cas9 nuclease. Creating AAVs capable of packaging larger payloads would simplify packaging not only in CRISPR research, but also in gene therapy applications. Purification protocols could be improved, too.
AAV applications may also be limited by the tropisms of the available AAV serotypes. Liu points out that if researchers are looking at a specific target in certain cell types, serotype screening may be necessary to determine the best fit. In some cases, a chimeric serotype may need to be created. “Over 100 naturally occurring AAV serotypes exist,” Liu notes. “We currently offer the most commonly used ones and supply AAV serotyping kits with ready-to-use particles to facilitate serotype testing.”
“Ideally, a designated optimized serotype for every cell type is desirable to reduce serotype screening,” Liu remarks. “No AAV serotype is 100% targeted to just one specific cell type.”
As a gene company, OriGene has the content for the entire genome in the form of validated expression plasmids. Any gene conforming to the payload limitations can be subcloned and packaged into an AAV particle using matching vectors, called AAV-ORFs. “A researcher can order PDL1-ORF,” Liu points out. “We have a verified PDL1 sequence, so we pop it out, drop it into the AAV, and package the particle. We do the same thing for researchers who use lentivirus.”
“We are enabling researchers who want to deliver genes using AAV but do not have sufficient cloning and packaging experience,” Liu relates. “AAV is only a tool. Let us, the tool maker, do our job so that scientists can do theirs, make discoveries.”
