The past five years have seen viral-vector-based gene therapies become a reality. To date, eight therapies have been approved by the US Food and Drug Administration (FDA) across three different types of viral vectors: adeno-associated virus (AAV), lentivirus, and herpes simplex virus.1 With 25 viral-vector therapeutics in late-stage development and another 120 in Phase II trials as of February 2022, the number of approved therapies is only expected to grow.2
The majority of early viral-vector-based therapeutics were developed within the context of rare diseases, either through direct administration to certain tissues or (in the case of late-stage or on-market oncology cell therapies) through ex vivo cell modification. In these contexts, small quantities of viral vectors were required, particularly as most therapies were still in the clinical stage of development. Now, with the shift beyond ultrarare indications, viral-vector manufacturing requires rapid expansion to be able to address these diseases in the commercial space. For example, some of the earliest viral-vector-based products targeting rare diseases required approximately 1,000 doses across development, access programs, and two years of commercialization.3 In comparison, the unprecedented demand and at-risk funding for COVID-19 vaccines enabled a ten- to 100-fold increase in production when adjusted by dose amount, with over two billion doses of the AstraZeneca viral-vector-based vaccine already produced.4
With a shift beyond ultrarare indications, viral-vector manufacturing requires rapid expansion to be able to address various diseases in the commercial space.
The increasing demand has required viral-vector manufacturing to rapidly evolve. The broader application of viral-vector-based gene therapies (for example, to more common diseases) requires higher yields and lower cost of goods (COGs). Large contract development and manufacturing organizations (CDMOs) have invested heavily in this space, with a number of large acquisitions in the past few years (Exhibit 1). While exciting, this rapid influx of money and new technology have not yet solved the bottlenecks and challenges of viral-vector manufacturing.
Unlike the production of monoclonal antibodies, viral-vector manufacturing is not standardized across the industry, with biopharma companies using different production systems and downstream processes. In some cases, different systems are even used across the portfolio of a single company.
The nascent nature of the technology also means that questions remain about the regulatory standards for manufacturing and quality control. While it is likely that several processes will converge and become standardized across the industry over time, key upstream processes will continue to act as important sites of differentiation between gene therapy companies. Making the right choices early will be critical for gaining regulatory approval and achieving patient access, and gene therapy companies will need to overcome key manufacturing challenges as they tailor their processes to their assets, while also making critical decisions at the asset and portfolio level that will allow them to leverage developments in viral-vector manufacturing to accelerate patient access (Exhibit 2).
Overcoming the current challenges of viral-vector manufacturing
There are three major challenges for large-scale viral-vector manufacturing that impact the different stages of viral-vector development (Exhibit 3). They are centered around the critical choice of production system; the ability to optimize downstream processing (DSP); and the development of standardized chemistry, manufacturing, and control (CMC) methods and quality assays.
1. No single production system is right for all assets
The choice among current upstream production systems remains a trade-off between flexibility, scalability, and quality (Exhibit 4). All three criteria are important to rapidly develop and commercialize a safe and efficacious blockbuster product. As no single platform currently delivers on all criteria, developers need to make an informed choice early on based on their confidence in the current construct, the size of the potential patient population, and the market landscape.
The flexibility to make minor changes to the construct (for example, gene variant, promoter) largely depends on the lead time needed to set up the production system. Transient-transfection systems allow the greatest flexibility, as they do not require prior cell line generation, which is why they are commonly used in early stages of development to rapidly optimize and test the lead candidate.
Scalability of upstream processes is essential for low-cost commercial production of viral vectors. This is a particular challenge, as in vivo viral-vector therapies such as AAV begin to expand into nonrare diseases. Transient-transfection systems rank low against this criterion, as adherent systems require a space- and labor-intensive scale-out, and transient-suspension systems become inefficient at larger volumes.5 Baculovirus expression vector (BEV) and producer cell lines are currently the most scalable and cost-effective systems with both being used in 2,000-plus liter bioreactors, although BEV achieves higher cell densities and around fivefold higher yields.6
Viral-vector-production quality is directly linked to the dose required to achieve efficacy, with higher-quality production enabling more doses per batch and potentially reduced likelihood of adverse events after systemic administration.7 One common quality indicator is a high ratio of full to empty capsids, with producer cell lines and BEV having superior results compared with transient transfection at scale. However, the insect cell–based BEV system produces viral vectors with lower infectivity due to nonmammalian post-translational modifications, leading to higher doses being required for efficacy.8 While purification can be used to remove empty capsids at a later stage, mammalian producer cell lines currently remain the most reliable option to produce high-quality viral vectors from the beginning.
At present, expertise and infrastructure in the gene therapy industry is still skewed toward transient transfection. All approved viral-vector in vivo therapies are currently produced in adherent transient-transfection systems. Once a production system is selected early in the development process, it is difficult to change without significant delays in the clinical or regulatory timeline.9 However, COVID-19 has led to increasing expertise and infrastructure and has accelerated innovation to scale viral-vector manufacturing. Viral-vector research and funding are currently directed toward the accelerated development of high-yield producer cell lines, which would enable higher-quality and more flexible viral-vector production at a lower cost.10 Hybrid systems such as packaging cell lines can combine some of the advantages of the different systems, but patents on many common capsids can limit the use to origin companies or partnerships. Novel systems that are being developed are trying to address all three criteria and could become interesting in the future; as yet, however, they remain unproven in clinical development.11
2. Individual optimization and low yields hold back downstream processes
Currently, lack of standardization and low yields are a challenge for DSP, leading to a high reliance on individual expertise and the skipping of quality-enhancing steps. And there is currently no standardized platform approach for commercial-scale viral-vector DSP (as is seen with monoclonal antibodies). Physical characteristics are highly variable between viral vectors and the degree of their accumulation inside cells versus media can differ significantly. This poses a challenge to developing platform approaches with predictable yield outcomes. Different options are available for primary recovery, clarification, and purification of viral vectors, including ion-exchange, affinity, and size-exclusion chromatography and flow-through processes, with different types and combinations used across companies. Consequently, a high degree of process optimization is still required for each product, yet low recovery from chromatography steps means that yields generally remain less than 50 percent.12 In contrast, monoclonal-antibody manufacturing achieves yields of greater than 90 percent through the use of highly standardized two- or three-step platform approaches.13 In addition, the larger size and greater sensitivity of viral vectors toward strong elution buffers result in higher losses per purification step compared with proteins. As a result, additional steps to increase product quality, such as enrichment for full capsids, are not commonly used across the industry at this point.14
Downstream processes will likely become more standardized over time as experience in the industry grows and platform approaches are developed for the most common viral vectors. While it is not yet clear whether standardization will result in a defined sequence of common purification techniques (for example, ion exchange), or whether newer or optimized methods will become common (for example, monolith chromatography, vector-specific affinity beads), there is a clear need for industry-wide sharing of best practices and guidelines to reduce variability.
3. Establish close collaboration with regulators early to avoid delays
Because of the industry’s limited experience with commercial, at-scale supply of viral-vector gene therapies to date, enhanced quality and CMC guidance are necessary. Regulatory authorities have issued, and continually expand, specific guidance for cell and gene therapies.15 The limited long-term safety data and the risk of inadvertent changes to the genetic makeup of patients have led agencies to err on the side of caution; agency requests for additional information regarding adjustments to CMC controls,16 good manufacturing process (GMP) certification of materials,17 and potency assays are notable examples of such delays.18 Potency assays pose a particular challenge, as their development takes significant time and is difficult to standardize; viral-vector potency relies on multiple steps, from infection, transcription/translation, to target-protein activity, and assays need to ensure they measure the most suitable end point. Additional considerations concerning time and spatial control of expression, cytotoxicity, and immunogenicity of the viral vector also require reference standards and clear regulatory guidance.
Another challenge is the adaptation of manufacturing from the preclinical stage to clinical and commercial standards, which has led to the delay of several clinical trials.19 Complexities include the necessity of using commercial manufacturing standards20 early on to avoid delays resulting from manufacturing-protocols adaption; the requirement for lengthy comparability studies to change or adapt manufacturing processes after initial preclinical or clinical results have been obtained;21 and ascertaining the cause of differences in clinical efficacy after manufacturing processes or locations have changed.22
Regulatory guidance will likely evolve over time and become more detailed as more products come to market. This could result in more specific guidance on potency assays and clearer paths for switching production systems as analytical methods for comparability studies evolve.23 FDA questions from a recent gene therapy hearing also show that regulatory authorities are assessing feasibility and need to regulate maximal viral dosages.24 The high variability between products, titering methods, and viral-genome load versus capsid exposure increases the complexity for regulatory authorities to set safe dose ranges. However, requiring companies to share existing data on empty-to-full capsid ratios would allow agencies to better assess links between dosage and toxicity, as at least one company has previously opted to remove empty capsids from their products in response to safety events.25 Last, as the demand and supply of CGT manufacturing continuously increases, there is a need for updated guidelines for the GMP segregation practices for multiproduct or multimodality facilities.
Regulatory guidance will likely become more detailed as more products come to market, which could result in clearer paths for switching production systems.
Unlocking opportunities in viral-vector manufacturing
Many of the current challenges in viral-vector manufacturing for gene therapy arise from the push to be one of the first to market due to the drastically reduced opportunity size for any followers with curative therapies. But more companies are starting to shift attention from first launches in rare diseases toward development in nonrare indications. Five strategic considerations can help them plan for viral-vector production.
1. Agile make-versus-buy strategy
A critical production consideration is whether to build manufacturing capacity in-house or outsource production to specialized CDMOs. Using external manufacturing platforms can allow a quicker ramp-up of production as well as access to technical and regulatory expertise and reliable GMP-certified facilities and materials. However, high demand means that start-up times for contract manufacturing can be more than 18 months.26 Increases in CDMO capacity, as a result of COVID-19 vaccine manufacturing, are expected to reduce this time and further support standardization of tech transfer processes. The use of CDMOs is particularly suited to smaller companies or early-stage development activities, with several examples of strategic partnerships with contract manufacturers.27 This enables access to expertise that can be valuable in a competitive market while avoiding large investments in manufacturing capabilities of a modality that may still draw questions about its applicability to a wider patient population. Nonetheless, several larger companies have built their own manufacturing capacity to develop proprietary production systems and enable lower COGs at scale.28 As the industry still awaits a one-size-fits-all production system, the jury is out on whether experience-based optimization by CDMOs or in-house manufacturing teams will bring the desired advantage in terms of cost and quality. Companies should leverage the speed of agile decision making in production decisions, as timelines for viral-vector products are compressed compared with other modalities due to the rapid pipeline evolution, accelerated regulatory pathways (for example, rare diseases), and the rapidly changing capacity and offerings across the CDMO market.
2. Smart capital deployment and planning for scalability
Companies investing in their own manufacturing capacity must make key decisions regarding the scale of investment required and choice of production platform. Modular production suites can provide some flexibility for the scale of production and allow expansion as key development milestones are met. The production suites can be largely set up in a platform-agnostic way, with the exception of adherent cell culture, as bioreactors, consumables, and downstream processes are mostly the same. Key differences between platforms exist during process development, where specialized capabilities are required to optimize the yield and quality of the specific viral vector. This is in addition to the separate production suites needed to generate master cell and/or viral banks or to develop producer cell lines. The executive suite must also set the size of the initial investment, which is both dependent on production yield and demand forecasts. Accurate prediction remains a challenge at this stage, with a global pharma company recently selling its large production facility to a CDMO only 14 months after opening due to process improvements and the evolving dynamics of the gene therapy landscape. This is an example of how capacity can shift between different owners to serve the overall ecosystem.29 Due to large differences in yield between production systems and continued advancements in process optimization, companies will likely need to be content with some level of uncertainty and risk when making their investment decisions. Viral-vector manufacturers could leverage digital tools to assist with this decision. For example, the use of a digital-twin system could allow manufacturers to model future demand and plan for manufacturing capacity accordingly.
3. Maintaining an end-to-end lens
A third key decision area involves end-to-end supply chain management for raw materials and viral-vector products. Several companies using transient transfection have been insourcing their supply of key raw materials (such as plasmids) in the face of industry shortages, long wait times, lot-to-lot variability, and high COGs.30 While suppliers have ramped up production, insourcing still brings both time and cost advantages,31 especially to firms with larger viral-vector portfolios based on the same production system. In addition to strategic decisions regarding vendors, companies also need to manage the supply of their own products. Viral vectors take about a month to produce and require long-term storage at ultracold conditions with maximum storage periods limited by regulatory authorities, posing a challenge to stockpiling excess supply.32 By maintaining an end-to-end lens that includes demand forecasting, stable formulations, and flexible, low-cost raw-material supplies, companies can optimize their supply chains and manufacturing capacities.
4. Leveraging digital to facilitate operational excellence
Manufacturers should consider how digital tools can further unlock operational challenges in production. For example, yield can be optimized with advanced analytics–based models, which can help manufacturers proactively identify and correct potential issues in yield. Long timelines for resolving deviations can be shortened with an analytics-driven recommendation engine built on historic deviation and underlying root-cause data. Separately, the industry-wide talent shortage could benefit from augmented and virtual reality. These tools can be used to train operators without removing the lab or SME (subject-matter expert) from production, whether for initial onboarding or for infrequent processes that require refreshers.
The industry-wide talent shortage could benefit from augmented and virtual reality. These tools can be used to train operators without removing the lab or subject-matter expert from production.
5. Getting ahead of innovation
Another key consideration for companies is how to apply new innovations to their manufacturing processes. For products in the development stage, this means identifying high-potential innovations that could bring significant advancements early on, such as high-yield adapted cell lines or novel production systems. While adoption of untested processes increases the risk of delays, especially in time-sensitive first-to-market scenarios, improved scalability or quality can become key enablers of accelerated patient access when launching second or third. Consequently, as the field of CGT is impacting more and more companies, successful ones will look for external innovations to apply to their own pipelines.
The outlook for viral-vector manufacturing
Manufacturing of viral-vector gene therapies will likely remain a key point of differentiation in the near future. Choosing the right production system, optimizing DSP, and developing standardized CMC methods and quality assays continue to be the most common challenges faced in viral-vector manufacturing. Based on their portfolio strategy and the proportion and development stage of in-licensed assets, companies need to balance between investing in the development of proprietary technologies and selecting the most suitable systems and methods for each asset. As more companies start to shift attention from first launches in rare diseases toward development in nonrare indications, thinking ahead and making key manufacturing decisions early on become critical for success.