Pilot plants represent a milestone in a drug’s commercial development, where drug sponsors first explore engineering, commercialization, clinical, regulatory, and process issues outside the relative safety of laboratory scale. And, as the test ground for a product’s ultimate manufacturability, pilot scale is where drug sponsors often return to de-bottleneck or to solve unanticipated production issues.
“Pilot plants are a bridge between preclinical and clinical, or between early and late clinical stages, for example after filing a New Drug Application,” says Greg Hoyt, bioprocess specialist at CRB, a facilities design, engineering, and construction company.
Traditionally, companies build pilot plants before they know if a drug will be approved. This uncertainty underpins critical decision-making on both pilot and manufacturing-scale capacity, the introduction of novel manufacturing technologies, implementing scaleup and scaledown strategies, and how to conduct knowledge and process technology transfer.
To these considerations Hoyt adds strategies for piloting in a multiple-product facility, flexibility in both production technology and scheduling, and the use of platform technologies. “It obviously helps if you’re working with similar molecules from one campaign to another.”
A host of issues arise simply as a result of moving from small to large scale. “You’ll almost certainly need to perform mixing studies, and if you’re using single-use technology there may be issues with incompatibilities in materials of construction, leading to unpredictable profiles for leachables and extractables,” Hoyt says. “Pilot scale may require using different pump technology, or changing from a benchtop centrifuge to a disk stack or continuous centrifuge.”
Moving from bench to pilot scale, and from pilot scale to manufacturing, entails nearly identical volumetric scaling—ten liters to 100 liters, and 100 liters to 1,000 or 2,000 liters, respectively. Yet the former, according to Hoyt, is more difficult, since “it requires changing equipment—not just the size but the type. It’s much easier to match columns, filters, resins, etc. between pilot and production.”
Operating at the interfaces
Knowledge transfer and capacity mismatches are often cited as challenges to successful pilot plant deployments, particularly at the interfaces of bench/pilot, and pilot/production. “One way to minimize upstream/downstream disconnects or capacity mismatches is to adopt continuous processing, as occurred at Pfizer’s Groton, CT, solid dosage facility,” says Maik Jornitz, CEO of G-CON Manufacturing. G-CON specializes in compact, modular, versatile pharmaceutical manufacturing facilities.
Jornitz refers to the 2013 project between Pfizer, G-CON, and GEA, which supplied Pfizer with a fully operational pilot plant just weeks after it was delivered. The plant used G-CON prefabbed cleanroom PODs (prefabricated cleanroom modules) and GEA’s OSD continuous manufacturing process.
“Oral solid dosage continuous processing can run at all scales, thereby avoiding the knowledge and operational gaps one typically sees when the site or processes are separated,” Jornitz adds.” Continuous bioprocessing could provide similar benefits to biomanufacturing, but it has not been widely adopted.
Since continuous processing is intensified compared with batch processing, and has a much smaller footprint, it is more conducive to modular cleanroom designs. “Podification of a continuous bioprocess is easier than for batch processes, as the volumes, equipment sizes, and footprints are much larger,” Jornitz explains. “The volumes of buffer and media are higher in continuous bioprocessing, but single-use technologies allow us to maintain large volumes outside of the cleanroom POD, and feed the fluids into the cleanroom via transfer ports. We are thereby able to keep cleanroom spaces compressed, which improves the economics and controls of running the facility.”
Continuous or semi-continuous processing can indeed help when matching scales, but Hoyt cautions that perfusion cell culture, and upstream operations in general, present their own unique issues. “Even capture chromatography in continuous mode is a challenge since, when loading a column over an extended run you have to think about the protein’s stability, bioburden control, and strategies for delivering many solutions without interrupting the main process flow. Continuous unit operations will definitely shrink your process if you do it right, but it doesn’t eliminate scaleup issues and may limit the scale of a process that is only as strong as its weakest link.”
Matter of definition
Not everyone agrees on the uniqueness of pilot scale within the context of approving medicines under GMP. Richard Welch, Ph.D., vice president of process and analytical development for the CDMO business unit at Emergent Biosolutions, believes that the traditional view of pilot plants as a separate, definable stage in pharmaceutical development is antiquated.
“Ask ten people to define ‘pilot plant’ and you’ll get eleven different answers. To me a pilot plant, whether operated under GMP or not, is where the small-scale models you’ve constructed during process development are finally put to test at large scale.” This holds whether the purpose of pilot-scale production is process validation or manufacturing clinical trial materials. “I view the pilot stage as a point along the development continuum where you determine if your production models are scalable.”
Emergent’s process development-scale operations are typical, ranging in volumes of 250 mL to 10 L, with “pilot” scale between 50 L and 200 L. The company’s approach to pilot-stage development is flexible: Scaleup occurs to both GMP- and non-GMP pilot plants, or in some instances in similarly sized clinical-scale manufacturing for production of clinical-grade materials.
Pilot scaling is effective, according to Welch, when equipment, operations, and support activities (e.g. analytics) between two subsequent stages match. When they don’t the mismatches must be identified and/or controlled. Exploiting non-GMP processes when applicable is a critical aspect of the company’s strategy.
“We scale up to 200 liters for non-GMP production before transferring to GMP,” Welch explains. “It’s far less expensive to run such initial scale up in a non-regulated environment, then to transfer the process scale-to-scale to GMP, or even to scale up to tenfold in GMP. This decreases the number of runs in suite, saves time, and greatly reduces the risk of technology and knowledge transfer. There’s a lot of benefit to scaling up during development.”
The roughly tenfold protein titer improvements observed during the last 15 years have altered the significance of scaleup, to the point where what was an investigational batch years ago may now produce as much material as production-scale batches did at the turn of the millennium. Yet this yield bonanza does not affect either the role or significance of “pilot plant,” according to Welch. The protein produced is used for preclinical studies and basic characterization, and the process knowledge gained is invaluable. “The only change I can see is that there’s a different end-volume target. There’s a great benefit in not needing to model a 40,000-liter bioreactor, either in development or at the pilot scale.”
The implications carry over for downstream operations as well. While it is generally easier to work with columns with diameters less than 1 meter, highly concentrated feedstreams can present their own unique challenges.