Evaluating antibody fragment purification using MabSelect™ VH3 resin

We packed MabSelect™ VH3 resin in a Tricorn™ 5/50 column and tested the dynamic binding capacity (DBC) at a 6 min residence time with purified VHH in triplicate runs. We used clarified Pichia pastoris supernatant with VHH at 4.9 g/L titer for purification runs on MabSelect™ VH3 resin in Tricorn™ 5/50 columns, according to the method in Table 1. We also performed purification runs using 0.5 M NaCl and 1 M NaCl in 20 mM phosphate pH 7.0 as wash buffer, with each run conducted in triplicate using a sample load of 21.6 mg (CV = 0.96 mL), which corresponded to 70% of QB_10%, We used the ANOVA test to statistically analyze results, for DBC and purification, considering yield and purity.

Table 1. Method for purification of VHH

*20 mM sodium phosphate,150 mM NaCl, pH 7.4

The QB_10% (DBC at 10% breakthrough) for VHH on MabSelect™ VH3 resin was determined by frontal analyses to be 32.1 mg/mL resin (Table 2).

Table 2. QB_10% results from DBC runs with VHH

*SD = Standard deviation

We carried out purification runs at load corresponding to 70% of QB_10%, with intermediate wash buffer containing either 0.5 M NaCl or 1.0 M NaCl. Overlay chromatograms are shown in Figure 1. We observed no product loss during washing─even using 1.0 M NaCl in the wash buffer.

Table 3 presents our yield, HCP, and aggregate (analyzed using size exclusion chromatography, SEC) results for elution samples. We found that a higher salt concentration of 1 M NaCl in the wash buffer gave better purity in terms of HCP clearance, when compared to 0.5 M NaCl (p = 0.015, at 95% CI by statistical analysis using t-test). The aggregate level is 2.0 ± 0.1% with 0.5 M NaCl wash, and 1.8% with 1.0 M NaCl wash. We obtained a higher yield using 1.0 M NaCl in the wash buffer than 0.5 M NaCl; however, this deviation is likely coincidental, and the yields are most likely the same.

Table 3. Summary of results from purification runs with VHH

*SD = Standard deviation
^†CI = confidence interval

Fig 1. Chromatogram overlay from purification runs with VHH using wash buffer containing 0.5 M NaCl (blue) and 1.0 M NaCl (orange), respectively. UV₂₈₀ (solid lines) and conductivity (dotted lines) traces are shown.

We packed MabSelect™ VH3 resin in Tricorn™ 5/50 and Tricorn™ 10/100 columns and tested the DBC at a 6 min residence time with purified Fab. We used clarified CHO supernatant with Fab at a titer of 1.24 g/L for purification runs, according to the method in Table 4. Additionally, we performed a wash-buffer screening study on the Tricorn™ 5/50 column. We used a load amount of 33.1 mg/mL, which corresponded to 70% of QB_10%. See Table 5 for a list of the wash buffers we used. Based on the outcome of the wash buffer screening, we carried out verification runs on the Tricorn™ 10/100 column with a subset of wash buffers. Following this, we analyzed the elution samples for yield, HCP, aggregate content, and ligand leakage.

Table 4. Method for purification runs with Fab

*20 mM sodium phosphate,150 mM NaCl, pH 7.4

Table 5. Intermediate wash buffers tested in buffer screening study with Fab

*End users must always ensure freedom to operate for the specific processes used for purifying their specific molecules

The QB_10% for Fab on MabSelect™ VH3 resin was determined to be 47.3 mg/mL resin. To test impact on purity and yield, we carried out purification runs at loads corresponding to 70% of QB_10%, screening different intermediate wash buffers (see Table 5).

Using standard high-salt intermediate wash buffer (0.5 M NaCl in 20 mM phosphate, pH 7.0), we observed some minor product loss in the wash step. High-conductivity Tris buffer and L-Arg containing buffer caused significant product loss (Fig 2). Other tested wash buffers did not cause product loss (Fig 3).

We determined the purity in terms of HCP reduction using different wash buffers, and plotted relative to HCP reduction using PBS as wash buffer (Fig 4). We found that

• Wash buffers with 0.5% Triton CG-110 or 1% Polysorbate-20 most effectively reduce HCP.
• Wash buffers using 50 mM caprylic acid give a similar level of HCP reduction to the 0.5 M NaCl wash buffer.

Fig 2. Overlay chromatogram from wash buffer screening study of buffers causing target loss. Least target loss is from standard wash buffer with 0.5 M NaCl.

Fig 3. Overlay chromatogram from wash buffer screening study of buffers not causing target loss.

Fig 4. HCP in Fab eluates from experiments screening different wash buffers relative to HCP amount obtained using PBS as wash buffer. Optimal reduction in HCP was obtained with 0.5% Triton CG-110 or 1% Polysorbate in the wash buffer.

We performed verification runs with Fab using five different wash buffers on Tricorn™ 10/100 columns:

• Three concentrations of NaCl: 150 (PBS), 250 and 500 mM in phosphate pH 7.0
• Two detergent additives: 0.5% Triton CG-110 or 1% Polysorbate 20 in phosphate pH 7.0

We analyzed these elution samples for yield, purity, and ligand leakage. Results from verification runs are presented in Table 6 and overlay chromatograms in Figure 5. We observed a minor UV peak early in the wash phase in the run using a 0.5 M NaCl wash buffer. SEC fraction analysis shows this peak mostly includes aggregated species, and it does not decrease the monomer yield compared to PBS (Table 6).

Using a 0.25 M NaCl wash buffer achieves comparable HCP clearance to 0.5 M NaCl, with no product loss during wash. The wash buffers containing detergent additives, as opposed to NaCl, give superior HCP clearance.

Table 6. Fab verification runs on Tricorn™ 10/100 result summary

*In phosphate buffer, pH 7.0.

Fig 5. Chromatograms from purification runs with Fab using wash buffer containing 0.15 M NaCl (blue), 0.25 M NaCl (orange), 0.5 M NaCl (red), 0.5% Triton CG-110 (green), and 1% Polysorbate 20 (purple). UV280 (solid lines) and conductivity (dotted lines) traces are shown. A different column with lower CV was used for the run with 0.5 M NaCl wash buffer, hence the lower load volume.