The processing of recombinant proteins from high cell density, high product titer cell cultures containing mammalian cells is commonly performed using tangential flow microfiltration (MF). However, the increased cellular debris present in these complex feed streams can prematurely foul the membrane, adversely impacting MF capacity and throughput. In addition, high cell density cell culture streams introduce elevated levels of process-related impurities, which increase the burden on subsequent purification operations to remove these complex media components and impurities. To address this challenge, an evaluation of mammalian cell culture broth buffer properties was examined to determine if enhanced impurity removal and clarification performance could be achieved. A framework is presented here for establishing optimized mammalian cell culture buffer conditions, involving trade-offs between product recovery and purification and improved clarification at manufacturing-scale production. A reduction in cell culture broth pH to 4.7-5.0 induced flocculation and impurity precipitation which increased the average feed particle-size. These conditions led to enhanced impurity removal and improved MF throughput and filter capacity for several mammalian systems. Feed conditions were further optimized by controlling ionic composition along with pH to improve product recovery from high cell density/high product titer cell cultures.

The safety of biopharmaceuticals is assured through the use of multiple steps in the purification process that are capable of virus clearance, including filtration with virus-retentive filters. The amount of virus present at the downstream stages in the process is expected to be and is typically low. The viral clearance capability of the filtration step is assessed in a validation study. The study utilizes a small version of the larger manufacturing size filter, and a large, known amount of virus is added to the feed prior to filtration. Viral assay before and after filtration allows the virus log reduction value to be quantified. The representativeness of the small-scale model is supported by comparing large-scale filter performance to small-scale filter performance. The large-scale and small-scale filtration runs are performed using the same operating conditions. If the filter performance at both scales is comparable, it supports the applicability of the virus log reduction value obtained with the small-scale filter to the large-scale manufacturing process. However, the virus preparation used to spike the feed material often contains impurities that contribute adversely to virus filter performance in the small-scale model. The added impurities from the virus spike, which are not present at manufacturing scale, compromise the scale-down model and put into question the direct applicability of the virus clearance results. Another consequence of decreased filter performance due to virus spike impurities is the unnecessary over-sizing of the manufacturing system to match the low filter capacity observed in the scale-down model. This article describes how improvements in mammalian virus spike purity ensure the validity of the log reduction value obtained with the scale-down model and support economically optimized filter usage.

No abstract

No abstract