Adapted from an interview with Phil Vanek. Joined-up thinking: success criteria for tomorrow’s viral vector production platforms. Cell Gene Therapy Insights 2018; 4(10), 873-877.
Demand fueled by clinical successes
The challenges and shortage of vector production reflect the fact that these materials and products are having such strong clinical success. Let’s celebrate the reason for this demand. To meet it, we need to industrialize and rethink how we transition from what has historically been carried out in translational clinical centers at smaller scale, using flasks and open systems, towards commercial production and methodologies.
Several areas must be addressed. First and foremost is the process itself: How can we improve efficiency in virus production? How can we change from planar surfaces, into 3D culture and suspension, to reach those higher scales that the clinical trials and commercial products will require?
And finally, like any other process, we need to look across the workflow and try to remove steps, complexity, and labor. At the same time, we must improve the use of materials, such as chemically defined media, produced under good manufacturing practices (GMP). Ultimately, we need to come up with cell lines that are highly efficient producers but also have the regulatory provenance that will be required.
Facilities designed to accommodate new platforms
We are an industry in transition, moving from translational centers, through clinical trials, where demand is increasing. And with the commercial successes we are now starting to see, as that scaling continues, we really must bring an industrial mind-set into the production of these materials.
Because we’re in transition, building facilities designed around today’s processes may not translate well into the technologies that are coming in the next 3 to 5 years. I think it’s essential that we keep flexible manufacturing, with open facilities that have reconfigurable footprints and floor plans to accommodate new platforms that are coming.
This may include new upstream technologies, such as bioreactors or other methodologies for scaling up, through to downstream purification processes. All that’s going to change in the next 3 to 5 years. Preserving that flexibility today for what will ultimately become a locked-down, buttoned-up process in 3 to 5 years, is probably the smartest approach.
Transition towards efficient GMP suspension systems
Planar culture surfaces have achieved very high titers and efficiencies, but one crucial problem is that they require substantial labor owing to manual steps such as feeding and passaging cells, as well as harvesting cells from those surfaces. As we look to scale up, these culture systems will also place increasing demands on existing facility space – and quite simply, moving from flasks up to multilayer flasks is not going to help us ultimately achieve the 1015 to 1016 viral particles that will be required, for example, in adeno-associated virus (AAV) manufacturing. So, there’s a need to move up to higher density culture systems.
Look back to what we have seen in the biologics industry, which moved from adherent cell types up into suspension Chinese hamster ovary (CHO) cells. That was the only way the industry could achieve the high concentration and productivity needed to support the commercial product.
The challenge we face as we’re exploring suspension platforms is that even cell lines modified to grow in suspension are still not achieving the productivity of cells grown on planar surfaces. Although there is a recognized need to shift to suspension culture in larger bioreactors, we haven’t solved the problem of manufacturing efficiency. As we are pursuing these 3D culture systems, there’s also an ongoing need to resolve issues around the biology, the vectorology, and all the process elements that will be influenced and changed to accommodate that scale up.
Considering the whole process when making changes
As we think about manufacturing and production, we must view it as a continuum, an entire process. Individual steps or unit operations all have ripple effects through the process. And it really starts upstream with the vectorology, the design of the producer construct, whether it’s the gene of interest or the other packaging and helper lines. All these components impact the overall efficiency of a manufacturing process.
It’s important that we approach our manufacturing strategy with that in mind: every time we change or shift one part it influences the next step in the process. For example, a move from planar surfaces into suspension culture or from batch to continuous manufacturing each affects the volume of culture that we require, the scale, and the facility where the culturing is done. It influences how we think about initially lysing cells, clarifying the raw material, putting it through the series of chromatography steps to clean up and purify the viral particles so they can be qualified and used downstream, all the way through to the final fill-finish packaging. Everything that we do has an impact. As we move to chemically defined cGMP, changes in culture conditions, larger bioreactors, or processing platforms downstream, there could be some consequences in the quality of material and efficiency of each step, upstream and downstream. And then ultimately the new processes must be validated.
I believe analytical tools will be critical for assessing any change we make in the process, but at the end of the day it’s really going to come down to better vectorology, a better ability to design and control the production of vectors. Then those improvements need to be combined into connected processes, such that every step is optimized. We must always think about the whole process, so we can achieve our goal of high quality, contaminant-free products with full GMP compliance and a regulatory track record.
Interest in standardization
Within the cell and gene therapy industry we talk about the requirement for standards and better control of raw materials and processes all the way through manufacturing. Therefore, certainly there would be value to a standardized or platform approach that could be shared and disseminated.
However, we’re asking, what does ‘good’ manufacturing look like? What is a quality vector? It is important to remember that many different viruses are being manufactured, although AAV and lentivirus are the main ones. Everything that’s designed into a virus to carry out the therapeutic effect, whether it’s a gene of interest that is going to have a modification effect or a CAR T antigen presented in a cell, may influence the ability to adopt a single standard approach or method.
So yes, I think there is an interest in a standard methodology or platform. But the challenge will be addressing the variability and differences in the vectorology that will impact the rest of the process.
Efficient process development with limited feed stream material
We hear often that this is a challenge. Process development is very expensive, especially when trying to push it to these larger scales. A 200 L AAV production run can be quite costly, so trying to optimize and manage the process changes, whether moving from planar surfaces to suspension culture, optimizing transient transfection conditions, or optimizing the approach of building packaging cell lines or establishing a stable producer cell line, will require what I call high dynamic range systems.
It will be important to choose technologies where process development and evaluation can be done at small scale, to allow scaling to larger bioreactors with confidence. Also, all the downstream chromatography and filtration applications will need to perform predictably across the scales of interest.
I think the key to developing efficient processes is going to be having technologies that scale very robustly, reliably, and most importantly, predictably.
How to prioritize improvements
I don’t think there’s a shortcut to addressing all the pain points in the process. One thing we hear often on the downstream side of manufacturing is the low yields from final filtration. We hear that yields in the 40% range are considered ‘good’ for lentivirus manufacturing using current methods. If I had a biologic and only recovered 40% at the end of the process, I would not consider myself very successful. But that’s where we are in the industry today.
When considering changes we need to prioritize the steps in the process that have the highest cost drivers (often the highest labor intensity), and then the ones that have the lowest recovery or efficiency in yields. At the same time we must start thinking about suspension-enabled cells, just to get to the higher culture density required to produce more viral particles in a particular batch. That’s going to be a critical element.
Other areas for improvement include downstream steps and products, such as affinity chromatography, ion exchange, and notably filtration steps. Each one will be slightly different and require the ability to look at the whole manufacturing process. In addition, it would be useful to have a capability or technology to help stabilize the labile viral particles during the production process, without influencing the viability or infectivity downstream.
So, there are several individual steps. From a process perspective the rule of thumb is go after what’s driving costs, what’s driving efficiency losses, and what’s contributing to high labor and touch points in the process. Those would be the areas I would encourage people to consider.
Addressing the talent gap
This gap is prevalent across all of cell and gene therapy. There are not enough people today that are very experienced in this sector, and there’s an element of us all learning as we go.
On the viral and vector side, substantial experience has been conserved from the bioprocess industry. Viral manufacturing, for the most part, looks and feels a lot like the biologics manufacturing process. So, there is a skillset being pulled over from that process with quite a bit of success. The big challenge we face, as I mentioned earlier, is that a lot of the processes being used today for manufacturing are coming from the translational centers. They had an acute need and built their own vector cores to manufacture relatively small batches. But these processes will not necessarily translate readily into industrialization or commercial-scale manufacturing.
To try to address these challenges, we’re going to need to bring more bioprocess individuals into the viral space. Now that there is a huge demand for viral vector, this is happening. The other focus must be on open communication and partnerships with the clinical and academic centers to support training initiatives. We as an industry need to establish centers of excellence where we can try technologies in the real world. The ultimate goal is to bring people in at the same time to train and learn from others. Also, it is important to build a network of key opinion leaders who have done this, done it at scale, to make sure there’s a community of likeminded individuals trying to solve this big problem.
Scalability, the science, the biology, the manufacturing, the quality – all have been done in other industries. So, we must borrow heavily from what we have already learned and be willing to admit we don’t know everything.