Protein A capture shortfall in a CHO antibody batch
A 2,000 L CHO fed-batch harvest contains a monoclonal antibody at 1.20 g/L. The capture step uses a 40 L Protein A column with a dynamic binding capacity of 35 g/L; it is run in two cycles. Pooled eluate is 2,100 L at 0.95 g/L. Step yields are: Protein A 83%, low-pH viral inactivation 96%, flow-through anion exchange 90%, and UF/DF 94%. After Protein A, host-cell protein (HCP) is 4,500 ng/mg mAb and aggregates are 7.0%. The anion-exchange step gives 2.0 log10 HCP clearance and reduces aggregates to 1.2%; UF/DF does not change purity. The release specification is at least 1.80 kg mAb, HCP at most 100 ng/mg, aggregates at most 2.0%, and final concentration 50 g/L. The final UF/DF retentate is 32 L. Manufacturing wants to release the batch and asks whether a single change - loading more harvest or widening the Protein A elution pool - can meet the mass specification.
1. Calculate the overall product yield, final mAb mass, final concentration and purity. Decide whether the batch can be released.
Read the worked answer
Starting mAb = 2,000 x 1.20 = 2.40 kg. Product after each step: Protein A 2.40 x 0.83 = 1.992 kg; viral inactivation 1.992 x 0.96 = 1.912 kg; anion exchange 1.912 x 0.90 = 1.721 kg; UF/DF 1.721 x 0.94 = 1.618 kg. Overall yield = 1.618/2.40 = 67.4%. Final concentration = 1.618 kg/32 L = 50.6 g/L, which meets the 50 g/L target.
Purity: HCP after Protein A is 4,500 ng/mg; 2.0 log10 clearance gives 4,500/100 = 45 ng/mg, below 100 ng/mg. Aggregates are 1.2%, below 2.0%. The purity and concentration specifications pass, but the mass is 1.80 - 1.618 = 0.182 kg short of the 1.80 kg specification. The batch cannot be released as tested; it must be reprocessed, supplemented from another batch only if validated, or the process must be changed.
2. Identify the yield bottleneck and calculate the Protein A yield or load needed to meet 1.80 kg.
Read the worked answer
Protein A is the lowest-yield step at 83% and dominates the shortfall. The required overall yield is 1.80/2.40 = 75.0%. The other three steps combined give 0.96 x 0.90 x 0.94 = 0.812. Therefore the required Protein A yield is 0.750/0.812 = 0.924, i.e. 92.4%, which is 9.4 percentage points above the current 83%.
Alternatively, keep the 83% Protein A yield and increase the load. Required load = 1.80/(0.83 x 0.812) = 2.67 kg. The column capacity is 40 L x 35 g/L = 1.4 kg per cycle, or 2.8 kg over two cycles. A 2.67 kg load is 95% of the two-cycle capacity, so it is feasible only with very little reserve for breakthrough and with a higher impurity load entering anion exchange. At 50 g/L, the final retentate volume would also need to be at least 1.80/50 = 36 L, not 32 L.
3. Recommend one change and state the evidence required before release.
Read the worked answer
The least risky single change is to increase the Protein A load to about 2.67 kg, preferably in three cycles of about 0.89 kg each (about 64% of dynamic capacity per cycle) rather than two cycles at 95% of capacity. This keeps the same 83% step yield and the same downstream yields, giving 2.67 x 0.83 x 0.812 = 1.80 kg. Three cycles add processing time but preserve a capacity margin and reduce breakthrough risk. Widening the elution pool to raise the yield to 92.4% would also meet mass, but it is likely to increase HCP and aggregate loading and is not supported by the current purity data.
Before release, demonstrate: Protein A breakthrough below the validated limit; HCP after anion exchange at most 100 ng/mg (target below 50 ng/mg for margin); aggregates at most 2.0%; final volume at least 36 L at 50 g/L; and consistent results across the three cycles. If the higher load causes any purity failure, the batch should be rejected rather than blended without validation.