Biomedical Engineering Reference
In-Depth Information
and the development work was concentrated on the third, polishing step,
where the multimodal anion exchanger Capto adhere was used.
The feed consisted of filtered CHO cell culture supernatant contain-
ing 2.7 mg mAb/mL. Sample volumes corresponding to 25 mg mAb/mL
bed volume were applied to the XK 16/40 and RTP MabSelect SuRe 2.5
columns. Five cycles, each including cleaning-in-place (CIP) with 0.5 M
NaOH, were run on each column, and the eluates were collected using an
UV watch function. mAb purification at large scale typically contains a
virus inactivation step at low pH after the protein A capture step, taking
advantage of the low pH of the collected eluate. This step was omitted in
this study. To match the buffer conditions of the equilibration buffer in the
subsequent Capto Q step, the pH of the collected eluates was immediately
adjusted to 7.6.
The pH-adjusted eluates from the five MabSelect SuRe runs were pooled
and applied to the Capto Q column in flow-through mode. The flow through
and part of the washing solution were collected and prepared for the Capto
adhere step by adjusting the conductivity and pH to match the conditions of
the equilibration buffer in the Capto adhere step.
All material from the Capto Q run was applied to Capto adhere in flow-
through mode. The flow through and washing solution were collected.
Samples were withdrawn for analysis at each stage of the purification pro-
cess. The amount of dimer and aggregates in the samples was determined
by gel filtration on a Superdex™ 200 10/300 GL column. Host cell protein
(HCP) concentration was determined using the CHO-CM HCP ELISA kit
(CM015, Cygnus Technologies). The concentration of leached MabSelect
SuRe ligand was determined by a protein A ELISA method using purified
ligand for the ELISA standard curve. The analyses were not optimized for
this particular feed and mAb. Three-step monoclonal antibody purifica-
tion was performed in parallel at two different scales. The overall yield
was 88% for both processes, achieving contaminant levels acceptable for
formulation. The three-step purification process is characterized by an
overall good yield, low ligand leakage from MabSelect SuRe, and efficient
contaminant and dimer/aggregate removal. It should be emphasized that
the process development in this study was limited, since the conditions for
the two first steps, MabSelect SuRe and Capto Q, are more or less generic,
while the final Capto adhere step required some evaluation of operating
conditions.
Both MabSelect SuRe columns were run five times each to investigate
the effects of repeated runs on column performance, as well as to gather
enough material for subsequent chromatography steps. The chromatograms
obtained were similar. The uniform performance was confirmed by the ana-
lytical results. Yields were stable during all cycles, and the contaminant lev-
els were comparable. The HCP level was efficiently reduced, and the ligand
leakage low, which is characteristic for MabSelect SuRe. The higher level of
ligand leakage in the first cycle, which was detected on both column types, is
