From The Editor | August 19, 2014

BPI Offers A Look Into QbD At Genentech

By Trisha Gladd, Editor, Life Science Connect

Trisha Gladd

Quality by Design is a term that ignites both excitement and apprehension among those in the pharmaceutical manufacturing and bioprocessing industries. While it offers many benefits to product safety and quality assurance, it is not often implemented because of the additional resources required. However, QbD relies on the principles of quality risk management, rather than one strict process. So once the concept is mastered, it can be adjusted as necessary for each individual project.  

Brian Horvath, Scientist and Technical Development Team Leader, Late Stage Cell Culture at Genentech, says the company’s QbD approach has been evolving over the past several years. It is used in a number of ways in the bioprocess areas—cell culture, purification, and formulation—which allows validation that the process as a whole reliably produces product with the desired critical quality attributes. Currently, Horvath and his colleagues at Genentech are using QbD for process characterization and validation work for cell culture processes that are near licensure. “We start with risk ranking and filtering, which takes all relevant data for each process step in the cell culture process and that gives us a basis for which variables are enrolled in characterization studies,” explains Horvath. “It’s really the documentation for what our rationale is in deciding which process parameters we need to test.”

In October, Horvath will present a case study at the BioProcess International Conference and Exhibition outlining how QbD was used during the characterization of a CHO-based monoclonal antibody. “The case study incorporates the best points of how we have evolved the QbD platform at Genentech,” says Horvath about the presentation. “Most of the discussion will focus on what experiments we do to support eventual license claims for each cell culture process step.  The more interesting points will be with the penultimate inoculum culture, what we call N-1, and the production culture, where in each step we do multi-variate statistically designed experiments, or design of experiments (DOE), around parameters that we rank as the most critical to affecting product quality and cell culture performance," which he states is fairly standard in the modern cell culture development world.

In the initial production DOE done at Genentech, there were eight parameters that were varied in the study together, and the main effects on any given output were measured. The outputs include all critical quality attributes (CQAs) in addition to cell culture performance indicators.  “With any DOE that has many variables like that, certain factors are identified as the most important and others are identified as not as important. A follow-up study for those important factors is then completed, which allows us to receive parameter estimates for their two-way interactions and on each CQA of the product,” he explains. “This is where I’m hoping we can stimulate some discussion because although we did perform an actual follow-up study to the initial DOE, that first study has projectability—embedded within it are 4 (or fewer) full factorials. These  “in situ experiments,” or the projectable factorials within the initial designs, aren’t technically the same as if you did the study on its own where the other four (or more) parameters would all be held at target set points.”

In addition, Horvath plans to cover other aspects of QbD strategy, such as a quantitative way to determine critical process parameters (CPPs) using a tool to rank order parameter effects based on knowledge of CQA acceptance criteria. “Essentially, I’d like to review our QbD strategy, those parts that may have evolved since others have presented on this topic, which includes the concept of projectability, adding in some other aspects of testing or surveys of the overall process that aren’t necessarily experimental.”

While he realizes a half hour may not be enough time to present and discuss every piece of the information he intends to cover, his hope is that both he and his team, as well as the audience, will benefit from the interaction among industry colleagues who share an interest in the QbD approach. And with the intention of BPI 2014 to be a place “where industry connects to share technical, scientific, and strategic innovations across all phases of bioprocess development,” Horvath’s intentions are right on target.

See you in October!