Spin It First: Solutions for Better Bioprocessing

Driven by the need to support the production of high-intensity cell cultures of higher volumetric scales, the demand for performance efficiency at each step of bioproduction has grown in tandem. When a manufacturer decides on what technology to use, multiple methods can be applied to help facilitate the harvest step of bioprocessing. The purpose of harvest is to separate the product of interest from the remaining cell debris and particulates left in the bioreactor medium following the upstream process. The drug substance is then run through a series of downstream clarification steps to remove any unwanted aggregate or particulate that may be left over from earlier upstream processes. Historically, many facilities used stainless steel centrifuges, which require cleaning via SIP and validation of such cleaning from batch to batch to assure regulatory compliance and sterility. In contrast, depth filtration is more scalable using a series of filtration steps to remove waste from wet biomass. Many modern processes bring unique challenges requiring a more flexible harvest solution. This allows for further facility optimization of cleanroom space and other efficiencies. The ratios of consumable to hardware costs vary between these two most common harvesting methods. Single-use technologiesPurpose of Evaluation This paper evaluated two technologies for direct comparison. The two harvest methods assessed were depth filtration—the most common technique utilized in the harvest market, and single-use centrifugation—a technology that has been leveraged primarily at the small to mid-volume range. Alternatively, stainless steel centrifugation has been used at large volumes, but this technology is different from single-use in that it requires steam-in-place (SIP) clean-in-place (CIP) systems and is always accompanied by extensive validation work. As a point of differentiation, stainless steel centrifugation is fundamentally less flexible and requires more upfront investment via capital expenditure (Cap-Ex) due to regulations regarding the cleaning and operation of these systems. Single-use centrifugation, much like any other single-use technology (SUT), is far more flexible because batch turnover is as simple as replacing the consumable and line sets. Single-use offerings are less Cap-Ex intensive due to their consumable-based nature and, therefore, incur more recurring operating expenditure (Op-Ex) costs when the drug is produced commercially. It is important to note that any form of centrifugation requires a secondary filtration step which typically follows the depth filtration method. A pure depth filtration harvest approach also has two filtration steps, where the product is passed through a specific-sized filter, then passed through a second that is sized even smaller to achieve proper separation. Figures 1a and 1b illustrate the concept behind the primary and secondary filtration steps needed for both depth filtration and centrifugation. We are proud to lead the charge in expanding the viability of single-use centrifugation with our launch of the Thermo Scientific™ DynaSpin™ Single-use Centrifuge. The comparisons made in this paper will focus on the implications of harvesting from production single-use bioreactor (SUB) between the sizes of 1,000 L and 5,000 L. While the product is viable outside of this range, this analysis focuses on comparing the results in a GMP clinical phase scenario where large amounts of product are required. Key features: • Significant consumable reduction at all volumes resulting in ~70% less depth filters and 78% liquid requirements (buffer, WFI, NaOH) • Extensive automation that meets 21CFR part 11 compliance and allows for recipe planning, data tracking, audit traceability, user access, and real-time process monitoring • Efficient separation which results in reduced burden on the filtration step allowing higher filter capacity, measured in Liters per meter squared (LPM2) • Poka-Yoke design that minimizes setup and takedown time and helps reduces the chance of user error • Push buttons walk away automation enabling more efficient operation including automation of priming, steady state operation, and shutdown & drain Feed Centrate Concentrate Centrifuge Feed Filtrate Secondary filtration Feed Filtrate Primary filtration Feed Filtrate Secondary filtration Figure 1a. Depth filtration process. Figure 1b. Centrifugation process. 2Methodology The theoretical modeling discussed in this paper was carried out with the BioSolve Process software developed by BioPharma Services Limited. This software has been used in multiple industry papers to evaluate the financial and operational implications of a given process and is also used by many drug manufacturers to model their processes in-house. The following information details the key inputs that were varied in the model and how that was subsequently used to generate key comparisons. Both depth filtration and centrifugation have a myriad of variables that could impact the performance of the harvest. Variables such as titer, peak cell density (PCD), packed cell volume (PCV), and viability can impact product performance and yield. In this model, all three are held constant to show how both technologies would compare under similar circumstances. Because this modeling focuses exclusively on the harvest step, upstream and downstream portions of the model are held constant between processes so as not to adversely affect the harvest step. For example, the 2,000 L scale assumes that both depth filtration and single-use centrifugation pull from the same type of bioreactor with identical seed trains. This also means that labor, material, and consumable costs are held constant. To avoid facility costs from skewing the data the cost of the facility itself (including suite buildout) was excluded from the model, though this would have an impact should one technology allow a manufacturer to produce a much smaller suite. Additional assumptions were made around the quantity of personnel needed for process setup and takedown, the time required to set up/takedown a process, and filter/centrifuge performance. Table 1 highlights the key assumptions made that had an impact on the outcome of the model. Upon building the harvest models in BioSolve Process and identifying a range in potential product performance, the sensitivity of said performance was tested by iterating on the expected range. This meant running the model with different filter capacities as called out in Table 1. In all, the modeling performed included ~50 iterations. The data produced from these iterations were then leveraged to produce the comparisons highlighted in this paper. Table 1. List of base assumptions used in the BioSolve harvest cost modeling. Depth filtration DynaSpin Single-use centrifugation Setup time 1.5 hours per filter housing rack for material movement, documentation, installation, hydraulic compression and inspection, and tubing management .5 hours per single-use rotor, per filter Takedown time 20 minutes per filter housing rack for tubing teardown and cleanup, and 10 minutes per filter for teardown and disposal .5 hours per single-use rotor, per filter Minimum personnel requirement 2 personnel 2 personnel Primary filtration performance range Depth filter: filter capacity of 70-120 LPM² Centrifuge: flow rate of 180-660 L per hour Secondary filtration performance range Depth filter: filter capacity of 140-240 LPM² Depth filter: filter capacity of 150-300 LPM² Additional assumptions Labor Costs: Calculations assume some pre-staging has taken place for both technologies. This means the filter housing racks or the DynaSpin unit are already staged in the harvest suite, but the consumables have not yet been placed in the equipment. Capital Cost: Capital costs only account for equipment, the cost of building out the suite is not included. Equipment