Antibody Titer Drop in a Perfusion Harvest Line
A perfusion CHO process produces an antibody at 1.2 g/L in the bioreactor. After a harvest line change, the clarified harvest entering the capture column shows a titer of only 0.6 g/L by Protein A HPLC, while the bioreactor titer remains stable at 1.2 g/L. The harvest line includes a tangential-flow microfiltration step with a 0.2 µm membrane and a hold tank at 2-8 °C for up to 24 hours. The capture step recovery is normal at 95%. Operators suspect the analytical method, but the same method gives expected values on a reference sample.
1. Propose a systematic approach to locate where the 50% loss occurs.
Read the worked answer
Sample at each unit operation: bioreactor, pre-filter, permeate, retentate, and post-hold-tank. Compare Protein A HPLC titers and also check for aggregation or fragmentation by SEC. If the permeate is low but the bioreactor is high, the loss is in the filtration step; if the permeate is normal and the hold tank is low, the loss is in the hold. Also verify that the analytical method is not affected by matrix changes, since the same method works on a reference sample but the harvest matrix may differ. Mass balance across the filter (permeate plus retentate versus feed) will show whether antibody is retained on the membrane or lost to precipitation.
2. The permeate titer is 0.6 g/L and the retentate is 1.8 g/L. Interpret this and recommend a fix.
Read the worked answer
A retentate titer of 1.8 g/L and permeate of 0.6 g/L with a feed of 1.2 g/L indicates concentration polarization or membrane fouling: antibody is retained on the retentate side rather than passing through. This is consistent with a 0.2 µm membrane operated at a flux that exceeds the critical flux, causing a gel layer. The fix is to reduce the permeate flux, increase crossflow velocity, and add a diafiltration or flush step to recover antibody from the retentate. Checking the transmembrane pressure and the membrane's molecular weight cutoff relative to the antibody is also warranted, since a 0.2 µm membrane should pass the antibody freely unless fouled. After adjustment, re-measure the mass balance to confirm recovery.
3. If the loss instead occurs in the 24-hour hold tank at 2-8 °C, what mechanisms would you investigate?
Read the worked answer
At 2-8 °C, antibody can be lost through aggregation, precipitation, or adsorption to surfaces. SEC would detect aggregates, and a check of the tank's wetted material and surface area would test for adsorption. A time-course sampling at 0, 6, 12, and 24 hours would show whether the loss is progressive, which points to aggregation or precipitation rather than a one-time adsorption event. If the loss is concentration-dependent, dilution or a formulation buffer change may help. The practical fix is to shorten the hold, add a stabilizing excipient, or use a low-binding tank liner, and to verify with a repeat run before releasing the process change.