Clinical/Commercial Manufacturing: BPI Theater @ BIO 2015

Tony Hitchcock (technical director, Cobra Biologics) 1:30–1:55 pm

The Expanding DNA and Gene Therapy Market

The gene therapy market has fluctuated dramatically over the past 20 years. But Hitchcock noted that over the past three years there has been a huge emergence of successful clinical outputs of gene therapy, with a large number of IPOs and significant funding. Even more interesting is the number of large pharmaceutical companies bringing gene therapy into their portfolios. Hitchcock pointed out that it is a varied market in terms of therapeutic areas — going beyond the original gene replacement indications into cancer, infection, neurology, oncology, and cerebral folate deficiency (CFD). One product is licensed in Liberia, but a number of other projects are coming out of phase 1 and some are approaching phase 3.

Market Projections: Hitchcock said that some analysts project a 20-fold growth in the market over the next 10 years. Eighty percent of that market is expected to be in oncology. Not counted within those numbers is the potential of combination therapies — particularly using immunotherapy in combination with other oncology treatments. Gene therapy “is not just a singular product; it’s very much a construction of delivery vehicles for therapeutic uses. So it effectively allows you to develop processes and manufacturing systems, including facilities, that are disconnected from one specific product. If you are invested in technical development, whether in manufacturing processes, analytics, or actually building facilities, you will not depend on the success of a single product. So from a CMO (contract manufacturing organization) perspective, that can be highly attractive. And if a drug development company has a platform for producing therapeutics for a number of different conditions, that should in theory reduce cost and allow an acceleration of those products to the clinic.”

The DNA and viral vectors markets are expanding. He described a number of research, therapeutic, and delivery approaches currently being investigated and related technical challenges in production and processing. Cobra offers both virus production and plasmid production. Hitchcock described significant challenges created by viral vectors, especially in scaling up adherent systems, for which there is large variability in productivities. Issues of virus containment affect facility design and the disposal of both liquid and solid waste. Cobra was an early adopter of single-use technologies for segregating such materials. Otherwise there is very little experience in product crossover and no acceptable limit of crossover for such products.

Purification requires one or two steps. A great advantage is the sheer physical size of the virus that needs to be separated from contaminants. But formulation, fill, and finish are challenging because products need to be stabilized and protected against aggregation. Hitchcock pointed out that a number of groups are invested in developing manufacturing processes for gene therapies, especially for those products that target rare diseases for which it’s necessary to sustain a supply of materials for a very small patient number. But such systems will be difficult or impossible to scale up to commercial production. So there is considerable scope for development and improvement of productivities and processing technologies.

Hitchcock said that customers present multiple and often proprietary vectors selected based on therapeutic approaches and delivery strategies rather than on manufacturing technologies and capabilities. This leaves open the potential to introduce platform processes and develop standardized approaches to these manufacturing systems. He pointed to challenges in scaling up for clinical production; developing analytical approaches for product characterization, stability, and functionality; and addressing questions from regulators.

Hitchcock concluded by noting that advances in working with plasmid DNA are critical to contract manufacturers, not only toward developing naked-DNA vaccine products, but also toward supplying and meeting the requirements of the viral vector market and adapting approaches to multiple therapeutic requirements.

Audience Questions: The first questioner asked what the American Medical Association actually requires as starting material. Hitchcock answered that guidelines are quite variable concerning what is recommended for early phase, late-phase, and GMP-production systems.

A second questioner asked for his thoughts on delivery through inducible gene switch mechanisms. Hitchcock answered that it is quite an interesting approach but not something that Cobra has been much involved with. He said that as long as a customer’s process and its limitations are understood, then as a contract manufacturer, Cobra can deliver to that quality system.

Watch the full presentation video on-demand (registration required)

Joe Wypych (general manager of the Seattle facility for CMC Biologics) 2:00–2:25 pm

Strategies in Large-Scale Manufacturing: Rapid Flexible Capacity Expansion with Novel Utilization of a Unique Single-Use Facility Design

CMC Biologics is a contract development and manufacturing organization with around 500 employees worldwide at three sites: one in Copenhagen, Denmark; one in Seattle, WA; and another being established in Berkeley, CA. The company has adopted single-use technologies and is commercially approved for global manufacturing.

Wypych began by explaining how CMC has been a pioneer in single-use (SU) technologies since the company’s inception. As an illustration of the flexibility single-use can provide, he described a 2011 installation of a 2,000-L SU bioreactor to increase capacity from 500 L. The change was implemented in seven months, and cell densities and viability matched those of the 500-L runs.

He then went on to talk about the “six pack” in Seattle. In May of 2014 a customer challenged the company to develop a plant to produce commercial supply for phase 3 clinical trials. The first task was to target capacity requirements based on using a 10,000-L reactor. At the time, CMC had two 3,000-L stainless steel reactors and two 2,000-L single-use bioreactors and did not have adequate ceiling heights to put in 10,000-L stainless steel reactors. So the company needed another way to achieve that capacity. Time lines were short for process characterization and validation, with the customer planning a biologics license application (BLA) submission late in 2015. So CMC worked backward from that target, creating a compressed time line.

The conceptual design was to install six 2,000-L single-use bioreactors. Capacity existed downstream, so the redesign could be focused upstream. Other challenges involved installing/replacing some utilities and the water- for-injection (WFI) system and not affecting ongoing operations for other clients. Using the existing downstream capacity, staffing levels needed to increase by about 50% and go to 24/7 operations. This gave CMC three to four months to bring 100 people on board and to get them trained and ready to start making material. The company also had to demonstrate comparability (viable cell density and titers) with material made in the 3,000-L stainless steel reactors, so it leveraged the expertise at the Copenhagen site for that.

After outlining the project time line, Wypych showed a facility layout, comparing the former and current configurations. With a goal to minimize the damage to the facility (and disruption of ongoing projects), the work was broken into segments. It began with demolition and clean-out of a seating area and worked toward the gowning area. A column-packing room and an expansion suite were added, along with a production suite and a harvest capture suite, as well as a new locker room.

The new production suites occupy about 190 m² of space, housing six 2,000-L single-use bioreactors, each pair of which has a single controller. Each of those three sets is harvested separately for clarification and concentration by tangential-flow filtration (TFF). After harvest, the materials are pumped to the purification area through a transfer line.

One result of the redesign was an improved personnel flow, maintaining a unidirectional flow throughout the facility (except in the glass wash/autoclave area). To minimize weight on the roof, I-beams were installed into the ground about 20 feet down to support a rooftop platform. Some of the sanitary lines were reinstalled and sloped a bit more than before to handle the increased amount of waste expected. A new air-handling system was also installed.

Within the building, CMC installed a different type of flooring than in the rest of the facility to guard against slips. And then the SU bioreactors were brought in and set up.

Advantages: Scalability is the first advantage Wypych emphasized. There were no significant scale changes from 2,000 to 12,000 L because the operation was still using 2,000-L reactors — six of them — and the process was not dependent on scale: Everything could stay on increments of 2,000 L. Costs and installation time lines are held down because such a facility is designed around building empty rooms with utilities, and equipment is then brought in. Cleaning validation is minimized, and processing is flexible using batch volumes. So if a client cannot predict at the outset what its commercial process will require, then that process can be easily scaled back later on. Finally, such a facility design is easy to replicate.

Wypych concluded by stressing that this sort of design is an excellent option for flexibility. Molecules can be changed out quickly, investment costs are lowered, and implementation times are shortened. This sort of design is easily adapted to large-scale manufacturing, whether for 12,000 L or 24,000 L. In the Q&A period, one question was about CMC’s total expansion strategy. Wypych explained that the company plans to use this format for all future projects: An empty warehouse can easily be converted to another six pack. Currently, CMC is installing a similar system in the Copenhagen facility, scaling up from an existing three-pack configuration.

Watch the full presentation on-demand (registration required)

Uwe Bücheler (senior vice president of biopharmaceuticals, Boehringer Ingelheim) 2:30–2:55 pm

From Transactional to Strategic Collaboration in Biopharmaceutical Manufacturing

Biopharmaceutical manufacturing is a dynamic market, with attractive growth rates and innovation creating business opportunities for pharmaceutical companies and contract manufacturing organizations (CMOs) alike. But many unknowns can complicate manufacturing development, so sponsor companies often choose to outsource manufacturing. Partners can help balance the risk for drug developers. The healthcare system is under pressure to lower costs, and biopharmaceuticals are on the higher end for cost of goods (CoG). Meanwhile, new markets are opening up around the world. And mammalian cell culture products are dominating the drug-development pipeline, making such technologies a major area of interest of CMOs.

The classical reason for outsourcing is that needed in-house capacities are unavailable, but an early stage product is too risky to justify building such capacity. Sponsors also outsource to access specific/proprietary technologies, platforms, and expertise. They also need to ensure a continuous supply of drugs to patients who need them, and CMOs can help them establish redundancies and make contingency plans. Many companies use dual sourcing for their blockbuster products to prevent drug shortages.

Other companies outsource their later-stage products while keeping early stage work in-house. Bücheler said that cost control is also an issue. Sponsors would rather not have the fixed cost of a fully owned plant for a single product. He pointed to strategic collaborations and partnerships as an additional element for outsourcing consideration. Some outsourcing relationships are transactional: one company paying another for services rendered. But others take a more strategic approach, not so focused on a single product or service but rather on the whole portfolio of services and collaboration elements that can generate benefits for both partners.

On the transactional side, cost and labor are the main drivers. It’s about capacity optimization or a lack of expertise. On the emerging partnership side are a number of models. Each partner might focus on a certain part of the value chain. So-called “virtual” companies might work with a number of partners to complete a full supply chain — e.g., a CMO, a contract research organization (CRO), and a marketing firm — while focusing on their own core R&D competencies. In strategic relationships, partners share investment risk and rewards. For example, increasing work on one project can compensate for decreased work on another in an overall portfolio. Bücheler said that co-investment and long-term collaboration can build trust between partners. Referring back to the transactional model, he said that such arrangements are finite, especially if a product ultimately fails. But strategic partnerships can continue for years.

Key success factors for CMOs, Bücheler said, include regulatory compliance and high quality standards. His company’s multiproduct facilities have been licensed since 1996. “We continuously work on improving our quality standards and systems,” he said. Each employee should have a “quality mindset” rather than leaving the subject up to a quality assurance (QA) department alone — and management plays a key role in that. Strategic partners must rely on one another, and neither one can stand still. “You have to update your technologies, your systems, and your processes.”

For example, Bücheler said his company is eliminating classical cleanrooms for aseptic filling and making the transition to isolator technology (despite the heavy investment required) to comply with current regulatory expectations. “This is what our collaboration partners expect,” he explained, “that we bring the operational excellence and flexibility to increase or decrease volumes.” His company needs to be able to take on new, nonstandard processes from clients new and old alike. A drug developer can develop a process on a single platform and then execute that platform. But a CMO must be flexible enough to transfer in and scale up its clients’ processes — while remaining competitive with respect to both cost and time lines.

Bücheler described his company, which is both a fully integrated pharmaceutical company and a CMO focused on the partnership model. Boehringer Ingelheim is the largest privately owned drug company in the world. To build a presence in the United States, it has acquired a cell culture facility in Fremont, CA; recently, it started GMP production at a factory in Shanghai as well. The latter is based on single-use technology. The company’s operations network also includes large-scale cell culture in Ingelheim am Main, Germany, and a flexible microbial fermentation facility in Vienna, Austria. Products in the microbial field, Bücheler said, don’t fit into platform technologies and tend to be smaller in scale. Finally, the company’s aseptic processing center in Biberach, Germany, won a facility of the year award in 2014 from the International Society for Pharmaceutical Engineering. He added that the company is looking to enter the medical device arena as well.

Watch the full presentation on-demand (registration required)

David Smith (vice president of global business development at Lonza) 3:00–3:25 pm

Cell Therapy Commercialization: How Do We Get There?

Lonza has been contract manufacturing for cell therapies since 2002, offering specific expertise in allogeneic stem cells. Smith’s first point was that all commercial-scale therapies using adult stem cells will require large amounts of cells – between 100 million and 1 billion cells per dose. To put that into perspective, an average (small) 5,000-dose therapy requires a trillion cells per year at commercial scale. The average lot size right now is about 10 billion cells, requiring production of almost 1,000 lots/year. So the industry needs to find manageable ways to scale up. As one example, he mentioned that TissueGene, Inc. is working on an osteoarthritis treatment to provide doses to 200,000– 300,000 patients. The latest estimate is that it will require over a quadrillion cells/year.

Required technologies need to be for adherent cultures, because the cells cannot be in suspension without being attached to beads, adding more and more beads as they go through passages. In the end, the cells must come off those beads alive and viable. As beads are added, large clumps of cells form in the bioreactor, so enzymes are needed to break them up – but too many enzymes can kill the cells. So the bead structure needs to release cells quickly and efficiently without killing them.

Smith noted that microcarrier technology will enable scale-up to these larger quantities needed for viable commercial cell therapy manufacturing. The rest of his presentation highlighted challenges of scale and processing technologies needed for cell therapies to succeed at commercial-scale quantities.

For the smallest of indications, about 10 billion cells can be grown in 10 stacks, producing only 100 doses. Using a high-density plastic, basically a flat 2-D surface, a maximum of about 100 billion cells can be produced, or about 500 doses per lot. Even for a simple indication, that would require making more than 500 lots of product each year – requiring QA to release more than 500 lots. Resulting batch records for one year would stack to a height of 40–50 feet – physically prohibitive.

Lonza’s move to bioreactor technologies required first assessing whether the company could produce enough product to compete economically with antibodies or other vaccines, and at a comparable quality. Smith’s illustration of this used data from standard models of mesenchymal stem cells (MSCs). A 100-L bioreactor can produce lot sizes of about 200 billion cells. A 200- to 500-L bioreactor could produce perhaps 1,000 doses. He reminded the audience that these are primary cells, which can double only so many times before they die. So a cap is reached between 250 and 500 L in a bioreactor, at which point expansion can go no further.

Growth Curves: Because therapeutic cells are grown in open suspension with agitation and flow around them, faster growth is possible than in 2-D cultures – taking a process down from fourteen to 21 days to nine or 10 days. “We can turn these lots over much faster whether we use MSCs or adipose or bone marrow. Also, as a contract manufacturer, we use all different types of bioreactors. Some cells like the configuration/agitation of certain bioreactors better than others.” As Smith explained, “The trick is really around the controller.” By comparison with two 200-L bioreactor trains, he said, just to get one trillion cells per year, “it would take about 600 Cell Factory tray systems; that means 600 plastic boxes that a technician has to pick up individually, change the media three times a week, passage and move them forward.” That would, he continued, require about 4,000 ft² of cleanroom space just for upstream culturing. That doesn’t even count the spaces for downstream, fill–finish, and visual inspection.

With the new yields, the cost of goods overall could reach a cost per dose that’s below antibodies, enabling clients to compete at a commercial level using bioreactors. He also explained how it is easier with a bioreactor than with cell factories to limit the particulates entering the product from plastic processing materials.

Cell Harvesting: Smith reminded the audience that time to harvest is limited for cell therapies – and that the drug product and drug substance are one and the same. “Once you start the process, and you formulate, you have less than three hours to finish your fill–finish, your formulation, your visual inspection. Because you’re using so much plastic, you are getting plastic particles in these products. So we had to find a way to limit the number of particles much better with a bioreactor. Particles are becoming the bane of cell therapy today with plastic processing (closed processing). Every time you change media, you have to weld a new bag onto the system. Every time you weld, you have particles.”

Filtration is not straightforward for these products. Sending them through too quickly creates shear that kills cells. They’re relatively large, so they can’t be pushed through a micron filter. So Lonza has developed techniques for high-volume, low-shear processing to decrease the number of particles that enter into the system down to a reasonable level without losing cell viability. “The result is large quantities and more important, we can get good quality systems. Now instead of dealing with 60 feet of batch records a year, you’re dealing with maybe five feet of batch records a year.”

Watch the video of the full presentation on-demand (registration required)

John Haney (senior project manager, Avid Bioservices) 3:30–3:55 pm

Designing and Implementing a New State-of- the-Art Single-Use Facility for Late-Stage and Commercial Production

Haney highlighted design considerations in the newest Avid biomanufacturing facility. The facility will eventually comprise three separate buildings in Tustin, CA, producing GMP material and performing analytical and documentation support. Since 2005 the facility has been licensed for commercial manufacture from 50 L to 1,000 L in single-use and stainless steelsystems. It has the analytical capability to support processes all the way to final product release.

Under the usual rubric of time, cost, and risk, Haney illustrated up-front decisions to be made for such a project. Should a company buy or lease property? Avid is leasing because funding a purchase is not how investors want their money to be spent. The next decision was whether to build from the ground up or start with an existing building and improve it. For Avid, that became a moot point when the building adjacent to the current facility became available. The next question was more complex: Did the company want to go with a stick-built or modular cleanroom? Weighing the pros and cons, Avid decided that the modular option would have a higher up-front cost but a lower lifetime cost and fewer maintenance requirements.

The company was under tremendous pressure to reduce the time line. Choosing a modular design saved three weeks out of the schedule overall because site preparation work was conducted concurrent with fabrication at AES’s Georgia facility. A fourth question was whether to adopt a traditional or a modern ballroom-style layout. Avid decided that the regulatory environment needed by its clients justified a traditional layout. It will have a larger, more expensive footprint, but management feels that it will be more robust from a regulatory point of view. A final decision was whether to go with single-use or stainless steel. From a building and an engineering point of view, choosing single-use technology eliminated the cost and time of testing and evaluating the large infrastructure that stainless steel would have created.

The new facility has (for example) no reverse-osmosis– deionization (RO/DI) or water-for-injection (WFI) systems, boiler plant, autoclaves, or parts washer. Utility needs are primarily for process gasses and electricity. Facility capacity accommodates three 1,000-L or two 2,000-L bioreactor trains, but the upstream suite is large, with ports to support client configuration requests. To be able to replicate the facility easily, it is designed with utilities and electricity, HVAC (heating, ventilation, and air conditioning), and other utilities to support building a second, identical suite.

Managing Partnerships: Haney acknowledged the companies and firms that are contributing or have contributed design, utility, and fabrication expertise to the project. He noted the importance of setting specific criteria. Some groups are “implicit partners that would otherwise fly below the radar, and you wouldn’t really think about them – but they are critical.” In a number of cases, designers have to accept and work around external time lines. One example was Avid’s need to accommodate Southern California Edison’s own procedures and scheduling. The review process for the permit package had to factor in the time it took for the city’s process of outsourcing engineering support. And easements and license agreements related to the surrounding property had to go through the landlord’s approval process.

His next slides showed the design layout. The drug-substance suite is about 9,700 ft², and the buffer and media preparation area is just under 3,000 ft². Color coding mapped the unidirectional flow and segregation of people and material. Every process area has its own air handler to prevent contamination between suites. A tubing pass-through facilitates closed transfer of aseptic material from harvest to downstream and from downstream into post-viral. Buffer and media are fed in through the wall. The monolithic construction has no open seams except for gaps under the doors, so it is airtight and designed for maintenance from a walkable ceiling without interrupting operations.

At the time of this presentation, the incubators were just being moved into the inoculum area over epoxy floors and connected to the environmental monitoring system. In the upstream area, Haney pointed to two 1,000-L HyClone single-use bioreactors that will be fed by a pair of 200-L Sartorius Stedim bioreactors – also single-use. The 14-foot-high ceiling provides room to expand to a 2,000-L bioreactor, and the space can accommodate six bioreactor trains.

The harvest area is open to the upstream processes to provide flexibility, but a removable wall can be dropped into place if a client needs to segregate upstream processing and harvest. A tubing transfer port allows aseptic tubing to be passed into the downstream area. The downstream suite is about 60-feet long with five process areas. It houses two 60-cm GE columns and five process stations. Equipment can be rolled out and a new downstream setup can be rolled in.

And at the time of this presentation, the facility had been constructed and air balanced, and the hope was to be able to schedule the first engineering thaw in the second week of July along with other work going on in the downstream suite. He detailed the time lines achieved and highlighted the modular cleanroom, “the shining star here.” Other elements still to be completed are buffer and media preparation areas and analytical laboratories.

This recording is being updated due to technical issues and will be made available at a later time.