The field of viral vector-driven gene therapy, incorporating both direct in vivo and ex vivo cell-based approaches, has enjoyed tremendous growth in recent years.
The key breakthrough of the first CAR-T cell therapies achieving market authorization worldwide (Novartis’s Kymriah® and Kite Pharma’s Yescarta®) has paved the way for further advances in the past 12 months, including the first AAV vector-driven in vivo gene therapy to receive US FDA approval (Spark Therapeutics’ Luxturna®).
In terms of disease indications, oncology continues to dominate with approximately 62% of all cell and gene therapy products targeted to cancer. However, orphan monogenic diseases are also well represented in the in vivo gene therapy space in particular, and increasingly, so too are larger patient population indications in key therapeutic areas such as cardiovascular and infectious diseases.
The gene therapy clinical pipeline has grown in step with rising valuations placed on individual gene therapy companies. It is now just over 2 years since the announcement of Gilead Sciences’ landmark acquisition of Kite Pharma for $11.9 billion and the trend continues unabated.
Of course, the speed with which gene therapy is driving towards a commercially successful future does present challenges, not least in the area of viral vector manufacture.
When considering the manufacture of any biological, whether it’s a monoclonal antibody or a gene therapy, there are certain key areas for consideration: ensuring raw materials, equipment and consumables are all compatible with GMP production is crucial; scalability of equipment and of each individual process step is equally important; and robustness and reproducibility are prerequisites for a commercially viable bioprocess.
However, the speed of evolution in gene therapy in general, combined with the fact that many product candidates are now on expedited development pathways (e.g., Breakthough Designation and RMAT in the USA; PRiME in the European Union) means that the bioprocess/CMC development window is effectively shrinking. In addition to this need for acceleration, gene therapy developers and manufacturers must contend with the reality that gene therapy bioprocesses are still derived to a large extent from the academic world, presenting problems both upstream and downstream in terms of quality, reproducibility and scalability. Not only that, but much of the current manufacturing equipment continues to be repurposed from the monoclonal antibody field. Although the effective repurposing of these tools is ongoing and novel solutions tailored to the specific needs of gene therapy production are emerging, this remains a work in progress.
Finally, gene therapy manufacture remains a comparatively highly complex field. There are multiple vector types, with multiple serotypes of these vectors (of AAV, in particular) used in the manufacture of products. Moreover, viral vector production currently involves the use of multiple cell lines. This lack of standardization stands in stark contrast to the more mature monoclonal antibody field, which has coalesced solely around Cho cells and a relatively very well defined, bioreactor-based upstream and downstream process.
This article will explore three vitally important areas in which strategic and technical manufacturing innovation is required to support commercialization:
Success in these three primary areas will be critical if we are to get past the current imbalance between supply and demand and effectively deliver on the enormous potential of gene therapy.
We can divide any viral vector bioprocess using mammalian cells into three main phases: upstream production, midstream processing, and downstream purification.
Upstream production typically starts with a cell thaw and expansion step, using either adherent or suspension culture. This is followed by a transition to the production phase, where cells are either infected or transfected to generate the viral vector.
Once the vector material is in the media, one proceeds to midstream processing, which begins with a vector material harvesting step. It may be necessary to conduct a lysis process to extract the virus from the cell, but equally, the virus can already be lytic and already present in the media, or indeed, it can be a combination of the two – it depends on the both the virus in question and on the particular program.
One may minimize the amount of host cell protein present through Benzonase® treatment before removing cell debris using stratification. Again, depending on the program and on the scale of production, it may be necessary to also do a concentration step at this point as well as a biofiltration through an appropriate buffer.
The next step is downstream purification where the main goal is to remove contaminants and obtain the pure product. That material is then concentrated prior to final formulation, final filtration and ultimately, fill-finish.
Of course, it is necessary to factor in testing of the entire process, including the amount of material available for this purpose. During upstream production, the volumes are large enough to mean there is no issue with testing but as you proceed down the production scheme, volumes become smaller and the material becomes increasingly precious – it’s important when devising a sampling and testing plan to keep this in mind.
This type of process usually works well at small scale but as demand increases, so too does the need to produce more material. One approach is to scale-out: carrying out the same defined unit operation multiple times. This is a good idea in principle because it is known that the unit operation works. Taking scale-up of cell expansion as an example, one might begin with T-flasks, then proceed to a slightly larger platform that is still in the flask format – layered flasks. At that point, one would scale-out by using multiple layered flasks. The same thing might apply with suspension cells, with the initial transition into shaker or spinner flasks followed by scale-out to multiple flasks. However, this approach only works well up to a certain limit, at which point a major bottleneck is encountered in the form of the amount of space and the number of manipulations required. Clean room space is limited – there is only room for a certain number of incubators. The amount of labor required is a further limiting factor.
Scale-out is therefore only viable for small- to mid-scale production. Beyond this scale, it is necessary to think about a different approach, i.e., scale-up. Scale-up uses the same concept of larger surface area as scale-out but requires a greater focus on the unit operations and then on the process itself.
For example, for scale-up for adherent cells, the same principle applies as for scale-out: one begins with T-flasks before moving on to a 10-layered flask stack. One can then proceed to a larger surface area again – 36-layer HyperStack® – and then do multiples of these. Coming to suspension cells, one might begin again with shaker or spinner flasks but then proceed to a small bioreactor (e.g., a rocking-bed or small disposable bioreactor) before transitioning once more to a mid-scale stir tank bioreactor.
This is a slightly improved approach in that it offers greater bandwidth. Nonetheless, one will still eventually encounter the same bottleneck of having insufficient space to conduct all the unit operations. Scale-up is therefore a viable approach only for mid-scale to perhaps the beginnings of large-scale production.
Novel technologies are emerging with the potential to enable genuinely large-scale production. These tools involve a scale-up approach but go beyond the traditional methods. For example, whereas in the past one might have gone from 10-layer cell stacks to HyperStack®, which would then be limited to around 20 or 30 per batch, it is now possible to go to a fixed-bed bioreactor offering a small footprint but a very large surface area of up to 500m².
Similarly, on the suspension cells front, huge strides have been made recently in bioreactor innovation. Devices are now available in a wide range of sizes – from 50 liters up to 2,000 liters – with footprints compatible with cleanroom capacities. This innovation is not particularly new – large bioreactors have long been used in the monoclonal antibodies field – but for the first time, demand for viral vectors has reached the level where their employment has become a necessity in gene therapy.
To summarize, there have been strong advances in upstream bioprocessing productivity over recent years. Today, the challenge is increasingly about optimizing midstream and downstream bioprocessing.
With regard to midstream processing, one must be cognizant of the larger volumes of viral vector material generated upstream: aware of the contaminants generated, for example, and the need to scale-up the clarification step accordingly. This requires in-depth knowledge of the flow rate used, the pressure, and the shear. Multiple filter trays – and multiple types of tray – may be required. These scale-related considerations all need to be factored in by the process development group.
The same applies to downstream purification. At small-scale, lower contaminant load means only a relatively simple purification scheme is required. However, it is important to remember that a more concentrated product comes with more concentrated contaminants. Volume is critical for downstream processing. One must factor in the type of media used during production – the protein composition of the chosen media plays a critical role in the binding of virus product to the matrices.
Typical virus size and properties, and elements such as pressure and shear must be considered in identifying the best chromatography and/or the purification scheme to use for a given application. Chromatography is generally recommended as a means of purification because of its robustness and scalability, which is of course crucial with larger volumes coming from upstream and midstream.
There are a number of variables and options available at this stage: One, two or three purification steps may be required to generate the material needed at the desired level of purity, depending on the eventual use of that material; bind/elute may be focused either on the product itself, or on the contaminants while the product flows through. These are critical development decisions impacting the robustness of the process.
Once the final product has been obtained, it’s important to understand the nature of that material in order to avoid massive losses at the final filtration step. (Again, depending on need and program, this may be followed by final formulation).
As demand for viral vector has increased rapidly over recent years, certain realities have emerged. Optimization steps are clearly required in order to arrive at a more robust, scalable process. Additionally, it is important to think outside the box when exploring methods to increase productivity, not to simply rely upon mass production. For example, harnessing novel fixed-bed bioreactors will help scale-up adherent cultures, potentially removing the need to transition or adapt your process to a suspension culture system, which can be challenging. Alternatively, using microcarriers in suspension might be considered, as indeed could the transition to completely serum-free suspension cells.
To conclude, while viral vector upstream bioprocessing has perhaps been the subject of the majority of focus and innovation to date, it is very important not to forget the midstream and downstream processes – to ensure that the whole is scalable, not just one or two component parts.
It is also worth noting that producing greater volumes of vector is not the only answer to meeting increasing demand. Further innovation and improvement in gene delivery and targeting should result in reduced quantities of vector being required in future. Ongoing efforts to optimize transduction efficiencies and/or delivery systems will hopefully lead to the same result. Finally, a number of companies are currently seeking to improve media, buffers and other components in order to better support cell growth and virus production.
Characterization and safety testing of the viral vectors used in gene therapy products essentially follows the same basic tenets as for all biologics, including identity, impurity, potency, and freedom from residuals of the production process: these are the fundamentals for assuring product safety and quality.
Identity focuses on demonstrating that the viral vector and its construct is what it is supposed to be, and that it remains so throughout the process.
Titer can be either biological activity, tissue culture infectious dose (e.g., TCID50) or it could be particle enumeration by qPCR.
Potency essentially describes how well the gene therapy product or the viral vector works. It can be based on a variety of different test methods, all of which relate to the mechanism of action or the expression of the transgene.
Purity is verification that the product is free from impurities and adventitious agents. The gene therapy vector is identified and possible contaminants such as related vectors or replication-competent vectors must be confirmed as absent.
Residuals are process intermediates and other holdovers from the process, such as host DNA, protein and enzymes that might be used in production.
As already established, the gene therapy manufacturing process is both complex and varied. There is no one-size-fits-all approach to viral vector production, and the exact same is true for characterization of a viral vector. Different vectors have unique characteristics. However, regardless of the specific approach used, safety and identity profile must be addressed for all components of the process. These include the cells that are used to produce the virus, the viral seed, helper virus or plasmids that are used for transfection or transduction of the cells, and the raw materials that go into the manufacturing process. All must be tested to ensure patient safety and to reduce risk in the manufacturing process as part of the overall quality of the product.
Throughout manufacturing, samples are taken for testing and as previously discussed, the amount of material available for this purpose varies throughout the process. At times, sample volume availability is quite low – a common challenge faced by manufacturers and companies going into a testing program. It is important that viral vector manufacturers as well as testing labs take this into consideration, particularly when designing methods and when setting up the sampling program. A well-designed sampling plan devised with this downstream testing in mind should be established during process development.
Testing of the cell banks used for production is a key component of the overall testing strategy. The cells form the foundation of any manufacturing program. It is worth noting that cell bank testing may take most of the total testing time available, and this should be accounted for upfront. There tends to be a greater focus on the vector or the downstream product testing, and it is often assumed that the cell characterization will simply occur – it can often be an unwelcome surprise just how much time it actually takes.
The following is a general outline for cell bank testing that would apply to any species:
There is now a constant flow of novel technologies coming to the field that enable faster testing and release for cell manufacturing.
Viral vector testing requirements depend to a large extent upon the vector in question. While all viral vectors must be tested again to confirm identity, titer, and to demonstrate purity, there may be additional safety considerations and concerns depending on the vector type and its application.
In the case of AAV, for example, the vector is also the final gene therapy product that will be put directly into patients. Testing is therefore more stringent to assure patient safety, particularly as it pertains to purity, freedom from residuals that might be part of the manufacturing and purification processes, and also freedom from recombinant replication-component virus in the final product. In general, AAV is considered relatively safe in this particular realm – it is a non-pathogenic and non-integrating virus. However, we must be able to demonstrate its safety.
By contrast, retroviral and the closely related lentiviral vectors are often not the final product, but rather are a raw material used in the transfection of cells – for CAR T cell therapy, for example. The focus of testing in this case is more on the characterization and safety profile of the viral vector. There are also additional testing considerations and requirements for the final cell therapy product, such as the pro-viral or transgene analysis. Again, though, identity, titer, purity and potency are all part of the testing scheme for retroviral vector batches.
As with cell testing, this space is benefitting from a steady stream of novel technologies that are enabling more agile and rapid testing and release of viral vectors. Additionally, as evidenced by recent guidance documents, the regulatory bodies are also building their understanding of the capability of these faster testing methods.
It is encouraging that both manufacturers and regulators are giving serious consideration to these technologies as viable alternatives to conventional methods. This will allow faster approaches to quality testing, which will be vital to the success of the many cell and gene therapy products with designated expedited development or accelerated approval pathways. It is strongly recommended that developers and manufacturers talk to their testing labs and also the regulatory agencies early in the process of planning the characterization and safety testing scheme. This is in order to understand the full range of rapid methods available, and to allow for consideration of alternative methods – either in parallel with or in place of conventional methods – that may ultimately help reduce time to commercialization.
Overall, testing should be thought of as a continuous process. As one progresses through the clinical development phases, the complex nature of the product as well as of the safety testing itself requires a variety of testing strategies for viral vector systems. These can be challenging to design, especially when trying to meet regulatory expectations. However, to reiterate, characterization and safety testing of viral vectors and gene therapy products is basically the same as for all biologics. If one designs or employs testing methods to assure the identity, purity, potency, and freedom from adventitious agents and residuals in the production process, then one should find greater ease on the path to commercialization.
Merck’s BioReliance® Manufacturing facility in Carlsbad recently dealt with the need to very rapidly get ready for inspections for commercial products as part of their BLA or European registration.
Traditionally, this site has worked with about 12–15 audits a year. Most of these audits were with biotech company clients. However, in the past 2–3 years, an increasing number of pharmaceutical companies have become involved in gene therapy and have visited the site. There has also been an increase in the number of visits from the regulatory agencies over this period. Carlsbad is in the state of California, which means the site had be licensed by the state even for the initial Phase 1 products. However, the FDA and EMA didn’t actually visit for inspections until the site was ready to manufacture a commercial product. At the time in question, three clients had either announced, or were expecting to announce shortly, a BLA submission.
Preparation involved not only reviewing the quality systems but the different processes involved for each of these three products, which included different types of viral vector. The initial expectation was for an 18-month window to prepare, and 6 separate workstreams were identified for further development.
However, the development of all three products accelerated, which resulted in the need to accelerate the site preparation program as well. The initial 18-month timeframe was reduced to 12 months.
The FDA visited in June of 2019, with the EMA visiting in September of the same year. Further visits from the regulators are expected, not only relating to the aforementioned three clients, but for other clients as they move towards future commercialization.
There are a few key learnings to share from this overall experience. It is strongly recommended to manufacturers to expect only around 12 months’ preparation time to achieve a state of commercial readiness. The longer the better, of course, but it is important to realize that Fast Track and PRiME designations will often accelerate development considerably. Bear in mind the requirement to fit in Process Performance Qualification (PPQ) runs, too.
One very positive impression received is that the regulatory bodies appreciate this is an emerging technology area – that everyone is in this together, and it’s important to work together in this spirit to ensure that patients can access these potentially life-saving therapies.
In the field of viral vector-based gene therapy where demand is far outstripping supply, there is a need within industry to scale-up rather than just scale-out. This is true whether the manufacturing is in-house or outsourced to a CDMO.
Clearly, initial viral vector process development is more scale-out than scale up, but in order to meet demand, it is certainly preferable for the manufacturer and the client to intensify their processes.
It has been established that there is a pressing need not only for scalability and reproducibility, but in conjunction with that, for appropriate testing to adequately measure the potency and safety of the product.
Finally, for those manufacturing facilities that are going to have an approved product, it is very important to approach preparations in a strategic manner – to develop workstreams to ensure all of the requirements are met.
As an example, the Merck facility at Carlsbad was able to bring in perspectives not only from elsewhere in the company, but also ex-FDA personnel who conducted mock audits along with all the regulatory oversight activities around the PPQ runs. It is obviously key to complete all these activities in time for regulatory inspections for a commercial product.
Contributions: All named authors take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published.
Disclosure and potential conflicts of interest: : The authors declare that they have no conflicts of interest.
Funding declaration: The authors received no financial support for the research, authorship and/or publication of this article.
Copyright: Published by Cell and Gene Therapy Insights under Creative Commons License Deed CC BY NC ND 4.0 which allows anyone to copy, distribute, and transmit the article provided it is properly attributed in the manner specified below. No commercial use without permission.
Attribution: Copyright © 2019 Merck KGaA, Darmstadt, Germany and/or its affiliates. Published by Cell and Gene Therapy Insights under Creative Commons License Deed CC BY NC ND 4.0.
Article source:This article has been created from a transcript of a webinar held by Merck featuring presentations by Dave Backer, Elie Hanania & Marian L McKee.
Webinar date: Oct 19 2017. Article publication date: March 19 2020.