On June 25, 2015, the Institute for Bioscience and Biotechnology Research (IBBR), the National Institute of Standards and Technology (NIST), and the University of Maryland (UM) put on its second Biomanufacturing Technology Summit, “Manufacturing Science Analytics: Predictors of Biotherapeutic Product Quality”, in Rockville, Maryland. McGuireWoods LLP was a sponsor of this important Summit and had Brian Malkin and TaeSoo “Sean” Kim on hand to interact with speakers and attendees.
Thomas R. Fuerst, Ph.D., Director IBBR, and William Bentley, Ph.D., Distinguished Professor and Chair, Department of Bioengineering, UMD, set the stage for the Summit as a place to discuss new bioanalytical testing for predicting the quality attributes of biological products including biosimilar versions of those products.
Then FDA’s Steven Kozlowski, M.D., Director, Office of Biotechnology Products, took center stage, where he said that in terms of analytics and biological products, there is always a need to strive for being better, challenging the current state of analytics that control quality as “not good enough.” For the same sponsor, changes in biological product manufacturing are driven by the quality control of comparability; for a different sponsor, the quality control goal is biosimilarity. Traditionally, FDA has thought about quality as a set of specifications, with changes that may be made either simultaneous to agency notification (changes being effected – 30 days), following preapproval supplements, or in annual reports. FDA has been concerned, Koslowski explained, that industry was confused with what were the “established conditions” in an application or license that represents the core chemistry, manufacturing and controls for an approved application (new drug application (NDA) or abbreviated new drug application (ANDA)) or license (biologics license application (BLA)). So last month, FDA issued draft guidance on the agency’s thoughts on “established conditions” in the guidance, “Established Conditions: Reportable CMC Changes for Approved Drug and Biological Products”. Here, “established conditions” is defined as “the description of the product, manufacturing process, facilities and equipment, and elements of the associated control strategy, as defined in an application, that assure product performance and quality of the approved product.”
Along these lines, Kozlowski described FDA’s evolving thoughts about the established conditions for biological products. Because the landscape of analytics to choose from continues to change, initial process or release analytics may change over time, resulting in a reordering of the risk ranking of quality attributes. As a result, it is critical to retain useful samples for comparative studies in anticipation of life cycle changes to compare products made with older and newer processes in view of the changing analytic tools. And it is increasingly important for manufacturers to integrate analytics into manufacturing to watch for potential changes in biological characterization (biocreep), as well as developing biosimilar products, where the analytics drive the required animal and clinical studies.
Regarding biosimilars, Kozlowski said that FDA has received over 80 meeting requests for 15 proposed biosimilar products, which have resulted in several filed and publicly-disclosed biosimilar applications and one approved biosimilar product. Kozlowski said that in terms of analytics, he believes that there is a need for a consortium of biosimilar companies to work together to help develop common and current analytics to help all of the biological product companies meet their goals more quickly.
The morning session featured additional presentations on real-time predictive models: bio-sensors and bio-predictors. Govind Rao, Ph.D., Professor of Chemical and Biochemical Engineering, Director of the Center for Advanced Sensor Technology, UM. Rao explained that it is critical to monitor biological manufacturing from start to final product; if not, “it would be like having a child and not monitoring it until it went to college.” Biosensors, such as oxygen levels, Rao observed, help to determine the window of optimal growth conditions for fermentation bioprocessing. Along these lines, Rao noted that there is a need to develop smaller-scale models that help to simulate ramp-up conditions for large-scale production for the final product.
Later, Patrick Swann, Ph.D. Senior Director, Technical Development, Biogen, discussed how different bioanalytics are utilized at various processing stages, including intermediate products or process-related impurities monitored, such as acidic isoforms that can limit shelf life, as part of quality-by-design. Alexion’s Kyle Zingaro, Ph.D., Development Scientist I, Upstream Development, then discussed how improved bioanalytic characterizations lead to accelerated process development and platform development with the goals of quality, robustness, raw material supply chain, scalability, productivity, and cost effectiveness. Zingaro described how the need for a better connection between biophysical quality and clinical safety and efficacy drives to a consensus development of a biological fingerprint, which will be increasingly important.
FDA’s Lieutenant Commander Cyrus Agarabi, Ph.D., Senior Regulatory Researcher, Center for Drug Evaluation and Research, rounded out the morning’s program. Agarabi described how FDA’s internal regulatory research, including biological product analytical testing, helps to support policy development and biotechnology review decisions. Agarabi encouraged industry to collaborate more with less compartmentalization of knowledge by sharing information about equipment and use for integrated bioprocessing systems to drive product quality in-line, on-line, and at-line with continuous manufacturing. As an example, nutrient analysis is often critical to affect cell viability and reliability, which can be integrated into bioreactors and autosamplers, including continuous sampling of media and cell membrane charges (indicating a cell is alive) to maximize cell harvest.
The afternoon focused on computational and analytical methods for predictive final product characterization. Michael Tarlov, Ph.D., Division Chief, Biomolecular Measurement Division, Material Measurement Laboratory, NIST, said that NIST is looking for more collaboration in the development of bioanalytical predictors for biological product quality. Steven Rose, Associate Director, Upstream BioProcess Engineering, MedImmune, described how MedImmune has been linking certain predictive cell culture scaling methods, e.g., pH and the carbon dioxide equilibrium to optimize inputs. Then Elizabeth M. Topp, Ph.D., Head, Department of Industrial and Physical Pharmacy, Purdue University, described how mass spectroscopy and other analytical methods may be used to test lyophilized biological products to obtain test and accelerated stability data, which has been conducted as part of a NIST consortium. Merck’s Marina Goldfeld, Associate Principal Scientist, described an innovative real-time measurement tool for glycerol, methanol, and biomass in a particular fermentation culture. And then Florian M. Wurm, Ph.D., Honorary Professor, Swiss Federal Institute of Technology and Founder and CEO, ExcellGene, described how Chinese hamster ovary (CHO) cells have been used for over 30 years for production of proteins in high quantities by making a selected cell with a desired gene “happy” through positive growth factors.
The Summit concluded with a robust panel discussion from the day where there was considerable discussion about how long it takes for an analytic method to become a standard and how there is room for collaboration for next generation analytics. Given the 12-year exclusivity for new biologics, even if there were sharing of manufacturing information now, Kozlowski speculated, it will only help the originator industry grow better. For instance, Kozlowski thought there could be a social-network-type way to share information without any one group feeling left out or some singularity. Zingaro thought that there were few new manufacturing processes to share but that there was some sharing going on with analytics with room to grow. As a “stretch goal”, Kozlowski thought that while the Agency is not at a point of saying clinical studies would not be required for market approval, in theory physical attributes could be linked to clinical outcomes. On a last point, it was agreed that better analytics at predicting quality would be a productive next step.