3.1.1. Performance at each step of end-to-end continuous process
for mAb purification
Process performance was examined for the end-to-end continuous process
incorporating upstream production processes with pool-less direct
connection, pooled low pH virus inactivation with automated pH control
and a total flow-through integrated polishing process in which two
column chromatography was directly connected to a virus filter. The flow
rates through the process were set to accommodate completion of batches
in succession to allow continuously utilizing the setups (Figure 1).
Batch 1 had a smaller volume but higher mAb concentration supplied to
the polishing step of 116.14 mL and 7.6 mg/mL (Figure 3), respectively,
due to air bubbles impeding the supply of harvest solution to the
capture process. For subsequent batches, the load volume and mAb
concentration were 315.29 mL and 5.1 mg/mL for Batch 2, and 317.97 mL
and 4.4 mg/mL for Batch 3, respectively. The pH- and
conductivity-adjusted mAb solution after the low pH virus inactivation
step was supplied to the flow-through two column chromatography
consisting of AEX and CEX columns directly connected to the virus filter
at a flow rate of 0.14 mL/min for Batch 1 and 0.38 mL/min for Batches 2
and 3. The flow rate were determined to adjust the batch cycle time as
about 24 hours. At these flow rates, the process feed had a residence
time for each column of 35.7 min for Batch 1 and 13.2 min for Batches 2
and 3. The mAb throughput for each column was 176.5 mg/mL-resin for
Batch 1, 321.6 mg/mL-resin for Batch 2, and 280.7 mg/mL-resin for Batch
3. At the end of applying the mAb to the flow-through two column
process, equilibration buffer (15 mL for Batch 1 and 75 mL for Batches 2
and 3) was applied to wash out residual mAb in the columns.
As the flow-through from the columns was directly loaded to the virus
filter, the filtration was conducted at the same constant flow rate as
the polishing chromatography step and the filtrate fraction was
consisting of the flow-through fraction followed by the buffer wash
fraction. At first, the filtration pressure of the Planova BioEX plotted
against filtration time (min) and throughput (L/m2)
shows pressure of about 0.073 MPa for Batch 1 and about 0.160 MPa for
Batches 2 and 3 for the filtration of equilibration buffer (Figure 3).
The pressure increased as mAb entered the filter, remained stable and
higher (about 0.077 MPa for Batch 1, about 0.175 MPa for Batch 2 and
about 0.170 MPa for Batch 3) during the flow-through mAb filtration,
then following a partial depressurization lasting several seconds due to
flow path switching, the filtration pressure decreased again. For
Batches 1 and 3, pressure fluctuation occurred during filtration of the
wash buffer, and this may be due to unstable supply of buffer from the
pump and introduction of air during path switching. Batch 2 showed a
slightly higher filtration pressure than Batch 3 due to a slightly
higher mAb concentration of 5.10 mg/mL compared to 4.41 mg/mL for Batch
3. For all batches, the filtration pressure remained almost constant
with almost no increase, even for throughput of more than 1000
L/m2 for Batches 2 and 3. The flow-through two column
chromatography process effectively removes impurities such as HCP and
aggregates (Figure 2), resulting in a very highly purified mAb solution
with high filterability. This demonstrates that the total flow-through
integrated polishing process adopted in the end-to-end continuous
process of this study is extremely effective for purification and for
filterability for virus removal.
Focusing on the virus filter performance, the flow rate through the
filter matches the flow rate through the columns and was 28 LMH for
Batch 1 and 76 LMH for Batches 2 and 3, and by volume-based throughput
(L/m2) and mass-based throughput
(kg/m2), the mAb throughput was 387.1
L/m2 and 2.9 kg/m2 for Batch 1, 1051
L/m2 and 5.4 kg/m2 for Batch 2, and
1059.9 L/m2 and 4.7 kg/m2 for Batch
3. In addition, the total filtrate volume including the wash from the
column, was 131.1 mL for Batch 1, 391.8 mL for Batch 2, and 393.9 mL for
Batch 3, and throughput for the filter including wash was 437.1
L/m2, 1301.0 L/m2 and 1309.9
L/m2, respectively for each batch.