This roundtable took place on Tuesday, 5 June 2018, in the BioProcess Theater at the BIO International Convention in Boston, MA. (Left to Right) Moderator: Dan Stanton, editor, BioProcess Insider. Panelists: Ohad Karniel (CEO and cofounder of Atvio Biotech, ISCT process and product development co-chair), Alan Moore (vice president and commercial chief for advanced therapies, WuXi AppTec); and Dr. David Brindley (University of Oxford and Harvard University)

Moderator Dan Stanton, with Ohad Karnieli, Alan Moore, and David Brindley

Moderator Dan Stanton introduced the second roundtable of Tuesday morning by commenting on the recent spate of approvals for CAR-T cell and gene therapies. With large biopharmaceutical and biotechnology companies devoting scientists and funds toward these next-generation therapies, contract manufacturing organizations (CMOs) also are investing in opportunities to expand capacity and expertise.

As the panelists introduced themselves, they raised two key points for discussion:

• The contracting model for advanced therapies is changing dramatically. Clients must establish both early and late-phase manufacturing capabilities to meet commercial requirements for scales and quality systems.
• The industry needs to accelerate automation and development of technologies to close what is still a huge gap between laboratory science and manufacturing at clinical scales.

The Kymriah Approval
The panelists were first asked to comment on the industry’s reaction to approval of Novartis’s Kymriah (tisagenlecleucel) coming shortly after Gilead and Kite Pharma’s Yescarta approval last year. Karnieli noted that advances in cell-therapy manufacturing have been incremental, but that approvals have brought increased attention to CAR-T therapies. Needed funds are being invested, and partners are stepping in to be part of this game. He said that five years ago, cell therapy developers were struggling to find larger partners to invest in their products, but these approvals are pushing the industry forward.

Brindley added that approval “has opened lots of doors and made people realize that there may be some substance to all the hype.” He also expects to see a second wave of gene therapies now, nearly 20 years after the failure of the first such product.

Moore commented that “we have been preparing for what we could refer to as ‘catastrophic success.’” For autologous production, substantial manufacturing gaps exist in technologies for supporting efficient GMP manufacturing and accommodating the expected numbers of patients. And investments in expanding facilities have exceeded the approvals so far. Recent successes have solidified recognition of the need to build development capabilities and unique quality control paradigms that will function at scale for the challenges of autologous products.

Outsource or Not?
Stanton turned the discussion to manufacturing decisions, asking whether big pharma companies won’t want to outsource such products.

Drawing from his experiences with Pluristem, Karnieli said that initially the company felt that a unique product needed to be manufactured in-house. On the other hand, contract development and manufacturing operations (CDMOs) face unfamiliar logistical concerns in moving manufacturing from hospital settings with limited numbers of patients to facilities that might need to make as many as 10,000 dosages of CAR-T therapies. How will products made at that scale get to hospitals for administration — to the right patients at the right time?

He also noted that although building a facility may seem efficient and economical, it can defocus a company’s efforts from resolving remaining gaps in manufacturing technologies. For that reason, he advises developers to find a manufacturing partner with experience in needed logistics and evolving regulations. With such a partner in place, a developed can focus on its clinics, its logistics, and on getting its product to market — rather than site development. “The right way is to focus on your target and let others do the manufacturing.”

Brindley agreed with Karnieli, saying that related decisions are driven by three things: capital, people, and complete chance. “In 20 years, pharma will want to bring manufacturing in-house because it’s a credible threat, and they want to be in control of their own operations. But over the next decade — maybe two — manufacturing will use hybrid models.”

Moore said that a number of companies still are very much insistent upon owning their manufacturing science and technology while controlling efforts to achieve the best cost of goods. But he also expects a hybrid role for CMOs in the long term. Outsourcing can bring access to a valuable platform technology, and partnering models are needed that will span geographies. He noted that some pharmaceutical companies are giving increased consideration to using another CMO for manufacturing in a second geography. In fact, his company has proposed serving as part of a regional system in which east-coast facilities could cover the east coast of the United States, and companies that want to build on the west coast can have their facilities there.

Karnieli stressed that if autologous cell therapy works, the whole model will have to change after the next decade or so. “It will go to the hospital. It will go toward the patient. Technologies will allow for decentralized manufacturing close to patients, dropping prices dramatically and making such therapies available. So although currently the focus is on getting to the clinic with as much manufacturing help as possible, he expects a change in 5–10 years toward manufacturing at the bedside.

Stanton wondered whether big pharmaceutical companies would engage in that sort of bedside manufacturing unit and how that would work. Brindley pointed to parallels in medical imaging in which a supplier such as GE provides a magnetic-resonance imaging (MRI) scanner, and a hospital buys the consumables and equipment from somebody else. Such a hybrid system for cell and gene therapies might include no role for CDMOs, but if a pharmaceutical company brings in a unit and checks it every couple of days, that could serve a CMO-like role as well as for securing consumables and QA/QC functions.

Moore pointed out that many cell therapies will represent a platform approach, but that not all platforms will be approved simultaneously. The expense of broadly distributed administration of approved cell therapies is a model that doesn’t exist yet from a regulatory standpoint. But there may be movement toward a central manufacturing environment in which CMOs would still play a logical role.

Moore believes in the potential and hope for bedside administration and processing of these cell therapies despite potentially changing the ways CMOs operate. “The exciting part is that if these kinds of pioneering products are successful, then there is a whole range of follow-on products directed at solid tumors, you name it, and a hyperpersonalization of cellular therapy. Some folks believe that we’ll see more rather than less personalization, and that could include producing allogenic cell therapies at the bedside.”

Logistics
Stanton asked whether logistical challenges posed by cell and gene therapies will continue to support a decentralized model. Brindley replied that there are two parts to the logistics issue: getting needed material (in an autologous model) to a processing facility — and then getting therapies back out to patients. A third bottleneck may be what some refer to as “that final fill and finish at the hospital/pharmacy/bedside.” Although using a courier to take some material from a bedside to a processing facility (and then return material to that patient), he expects final hospital fill–finish steps to be performed in a hospital pharmacy rather than an operating theater — which he said will be challenging enough “because most pharmacists have never seen a gene therapy.”

Karnieli said that the process looks easy in these early days: A certified nurse has been trained on how to thaw the product. A CRO watches over each patient to make sure that he or she receives the therapy. “But imagine that the pharmacy is three blocks away. Most subproducts have a shelf life of about two to three hours after thawing if you don’t put them in culture media. That sound like plenty of time. But what if a pharmacy thaws it, and the physician is inside the operating room — but the patient is too ill for the therapy to be administered that day? Even now, based on my experience, quite a few batches almost don’t make it to target.” So an important task for process development is to simplify transfer and delivery, perhaps by placing product in a device such as a prefilled syringe so that it won’t be thawed until it is at the bedside.

Moore agreed with the importance of considering the logistics upfront. Shifting the time of collection from a patient requires shifting manufacturing schedules. A QC team that’s waiting for a package to arrive at 4:00 pm might need to be rescheduled. So that presents a resourcing challenge.

“We’ve started to look at relationships with folks that are good at logistics and distribution,” Moore said, “people who have mastered the supply chain and can provide real-time tracking information. We’re working with the companies that have the tracking software systems. That’s all going to be the requisite going forward, particularly if you’re trying to optimize the throughput of a manufacturing facility. Getting the package to the hospital is not so much the challenge, but having the hospital or a physician or a nurse-friendly presentation of the therapy is and developing a cryopreserved medium that is appropriate for a certain product so that you can thaw it and place it in a bag for delivery.” Those steps will become simpler as technology advances, and alternative methods (such as for transduction of nonviral cells) will add different levels of complexity for manufacture, transport, and delivery of materials.

Open and Closed Systems
An audience member asked about the impact of closed systems on both manufacturing and delivery of cell therapies. Panelists responded by tying in the need for automation with closed systems — and current progress toward achieving both of those goals.

Karnieli commented that closing the system will limit overhead costs, reduce risks, and save time — even if only by removing contamination risk. It will thus lower infrastructure needs. Closing the system is the first critical step in development of these products. Automation then will bring in sensing and artificial intelligence (AI).

Brindley agreed that various types of AI and machine learning will be applied in future manufacturing suites and logically would extend to the hospital. He added that such a technical solution might conflict with reimbursement processes: At bedside care, every product and procedure has a reimbursement code. So he thinks that future processes will offer incentives for people to use a certain number of devices over a certain amount of time and in a certain class of environment to fit into hospital planning cycles.

Capacity
Stanton asked whether the industry has adequate capacity for cell- and gene-therapy products. Brindley identified two related issues: an overall lack of knowledge and a possible dearth of available capacity for process development, technology transfer, and CGMP and for meeting needs after that. Moore noted that some familiarity is emerging, at least in the CAR-T space where there is some concept of a similar platform for producing autologous cells. But the reality is that the field has never before enjoyed such a level of interest, funding, or technology development and advancement as currently — and all that is happening very quickly. “There is a real need for folks who have knowledge of the intricacies of aseptic processing and cell therapy products, particularly when working with induced pluripotent stem cells (iSPCs) that require a challenging, complex process to achieve the desired cell population.” He sees a shortage of resource capacity as a serious constraint and emphasized that much more training is needed in GMP processes specific to cell therapies.

Karnieli emphasized that every product is unique. So another large gap may be in the failure to plan for these to be dramatically different therapies. He also emphasized the need for increased training. His company trains groups internally and within in its global network. But it also offers training as a service. He is seeing such training needs increase, but says that “it’s going to take time until we fill that gap. Until then, it’s adding a lot of cost to the industry to recruit people who do have the expertise.”

Moore mentioned the importance of attracting employees from the pharmaceutical industry. He said that there is quite a lot of excitement in this space from knowing “that there is a patient on the other end of the product that you’re making, not just a dusty warehouse somewhere.” He also said that “the pace of training is very rapid when you’re trying to get a cell therapy rapidly into the clinic. You have to learn a lot, and you have to drink from a fire hose, so to speak. It’s evolving. It’s going to be some time before the industry has adapted to the unique requirements of cell therapies.”

Supplier Support
Stanton asked how the panelists felt about their relationships with vendors. Karnieli replied that “suppliers are in the space and willing to innovate. They want to push forward and get a good position. The gap is in understanding what we need. That is, we don’t know what we need, and they don’t know what we need.” He emphasized a need for innovation to grow from the bottom up. “My interest is in developing technologies for this industry, so being with a CMO allows me to gain that hands-on experience, to understand GMP issues, and to talk with engineers. Scientists don’t know how to talk with engineers — how to explain or even envision what they want.” But he is encouraged by the progress so far.

Moore noted how the Corning HYPERstack platform is allowing WuXi to produce viral vectors flexibly, yielding up to 250 L from an inherent system. Flexibility and collaborations are key because most people in the gene therapy space don’t need a 2,000-L reactor yet. “Bioreactor companies have awakened. They recognize the potential here. They’ve been listening to the expectations that the biomanufacturers have and the requirements for efficient production. Single-use systems are amenable or are morphing to accommodate the needs of cell therapies.” Tool companies also are continuing to innovate, spurred by recent commercial successes.

Brindley, however, pointed to remaining obstacles in development of tools and technologies, noting that there are long waiting lists for much of the equipment. He advises companies to repurpose tools and technologies wherever they can so as to prioritize efforts toward codevelopment of what is truly not adaptable. Moore agreed. “We don’t appreciate the complexity of establishing a new production line for bags and what kind of investment that takes. We’re seeing a lag, certainly,. But what’s encouraging is that the dialogue is started, and we’re starting to talk about the numbers of units we expect to need in the future.”

Watch Online
Watch the full interviews online at www.bioprocessintl.com/BIO-Theater-2018.