Mitigation of chromatography adsorbent lot performance variability through control of buffer solution design space.

The separation of undesired product-related impurities often poses a challenge in the purification of protein therapeutic species. Product-related impurity species, which may consist of undesirable isoforms, aggregated, or misfolded variants of the desired monomeric form of the product, can be challenging to remove using preparatory scale chromatographic techniques. When using anion exchange chromatography to remove undesirable product-related impurities, the separation can be highly sensitive to relatively small changes in the chromatography operating conditions, including changes to buffer solution pH, buffer solution conductivity protein loading, and operating temperature. When performing difficult separations, slight changes to the chemical and physical properties of the anion exchange adsorbent lot may also impact the separation profile. Such lot-to-lot variability may not be readily measurable by the adsorbent manufacturer, since variability can be highly dependent on a specific protein separation. Consequently, manufacturers of chromatographic adsorbents may not be able to control adsorbent lot to lot variability tightly enough to prevent differences from occurring when performing difficult product-related separations at the preparatory scale. In such cases, it is desirable to design a chromatography step with a control strategy which accounts for adsorbent lot to lot variability in the separation performance. In order to avoid the undesired changes to process consistency and product quality, a proper adjustment of the column operating conditions can be implemented, based on the performance of each adsorbent lot or lot mixture. In this work, we describe how the adjustment of the column buffer solution composition can be used as a design space based-control strategy used to ensure consistent process performance and product quality are achieved for an anion exchange chromatography step susceptible to adsorbent lot to lot performance variability. In addition, a "use test" is described that can be employed to determine the optimal buffer solution compositions for different anion exchange adsorbent lots based on the retention volume of the therapeutic protein during a gradient elution.

Efficient on-column conversion of IgG1 trisulfide linkages to native disulfides in tandem with Protein A affinity chromatography.

Protein trisulfide linkages are generated by the post-translational insertion of a sulfur atom into a disulfide bond. Molecular heterogeneity was detected in a recombinant IgG(1) monoclonal antibody (mAb) and attributed to the presence of a protein trisulfide moiety. The predominant site of trisulfide modification was the bond between the heavy and light chains. The trisulfide was eliminated during purification of the IgG(1) mAb via a cysteine wash step incorporated into Protein A affinity column chromatography. Analysis of the cysteine-treated mAb by electrophoresis and peptide mapping indicated that the trisulfide linkages were efficiently converted to intact disulfide bonds (13% trisulfide decreased consistently to 1% or less) without disulfide scrambling or an increase in free sulfhydryls. The on-column trisulfide conversion caused no change in protein folding detectable by hydrogen/deuterium exchange or differential scanning calorimetry. Consistent with this, binding of the mAb to its antigen in vitro was insensitive to the presence of the trisulfide modification and to its removal by the on-column cysteine treatment. Similar, high efficiency trisulfide conversion was achieved for a second IgG(1) mAb using the column wash strategy (at least 7% trisulfide decreased to 1% or less). Therefore, trisulfide/disulfide heterogeneity can be eliminated from IgG(1) molecules via a convenient and inexpensive procedure compatible with routine Protein A affinity capture.