January 15, 2016 (Vol. 36, No. 2)

Can a Workflow that Is Automated Provide Comparable Results?

Buffer preparation and exchange are es-sential steps in the development of biopharmaceuticals. These processes require many manual steps that are a significant limitation during formulation development.

Current approaches for small-scale buffer exchange include dialysis, desalting columns, tangential flow filtration, and centrifugal filters. While these devices are all relatively easy to use, each requires many manual steps, and none of these methods is ideal for exchanging a large number of formulations at small volumes.

Freeslate’s FormPrep system addresses these limitations by automating buffer preparation, buffer exchange, and sample concentration in a single, easy-to-use system. In one day, FormPrep can prepare up to 12 different protein formulations at the 0.5–8 mL scale.

FormPrep prepares up to 12 formulation buffers, which are subsequently used to create 12 unique protein formulations after buffer exchange. This system uses a novel ultrafiltration/diafiltration (UF/DF) technology to perform buffer exchange and concentration on an automated platform. To be suitable for biologic formulation development, a buffer exchange process must provide excellent material recovery and not impact short- or long-term protein stability.

Here we compare this automated buffer exchange process to traditional methods, by testing multiple proteins immediately after buffer exchange and after accelerated stability.


Study Design
Four confidential biopharmaceuticals, mAb α, mAb β, mAb γ and one enzyme, over a concentration range of 1 to 50 mg/mL, were used to compare the effects of buffer exchange by dialysis, centrifugal filters, and FormPrep. A stability study at 40°C was conducted (Figure 1) to assess the impact of buffer exchange on the long-term stability of each protein. All four proteins were buffer exchanged and then analyzed immediately after exchange (T0) and at two, four, and six weeks by a variety of analytics.

Each sample was dialyzed against 4 liters of buffer twice at room temperature (4 hours) and once overnight (14–16 hours) at 40°C.

Centrifugal UF/DF
Centrifugal filters were used to buffer exchange per the manufacturer’s instruction. Each sample was concentrated to 50% the initial volume and then diluted back to the initial volume for a total of five cycles.

FormPrep was used to buffer exchange each molecule using a concentrate-dilute procedure with gentle orbital mixing. Buffer exchange by FormPrep was accomplished by repeating cycles of 50% concentration, followed by dilution with buffer to the initial volume until 99% exchange was achieved.

Figure 1. Study design for comparison of buffer exchange methods.

Processing Times, Recovery, and pH Accuracy

Dialysis and centrifugal UF/DF filters are widely used methods for buffer exchange in formulation labs. Dialysis is gentle on proteins, but is slow and does not allow for concentration. Centrifugal filters are not as gentle on the protein, but are faster and have the additional capability of concentration.

Both methods provide high protein recovery and achieve satisfactory exchange efficiency, but require significant hands-on time. To assess if FormPrep could provide a suitable alternative to these methods processing times, protein recovery and pH accuracy were compared (Figure 2).

Processing times for FormPrep are comparable to centrifugal filters for the proteins at 1, 5, and 20 mg/mL. mAb β at 50 mg/mL required approximately five hours to reach 99% exchange by FormPrep compared to three hours for centrifugal filters. While FormPrep required a longer processing time it was fully automated and did not require any hands-on intervention. Protein recoveries for all methods and tested molecules were above 91% with FormPrep recoveries above 95% for all proteins tested.

The pH of each protein formulation post-processing was compared to the target pH of the formulation buffer. Centrifugal filter and FormPrep formulations were all within 0.10 pH units of the target pH. In contrast, when proteins at higher concentrations (mAb β at 50 mg/mL, enzyme at 20 mg/mL) were buffer exchanged by dialysis, the pH of the final formulation were 0.28 to 0.37 pH units from target. These results indicate that FormPrep provides comparable results for recovery and pH accuracy compared to centrifugal filters and dialysis.

Figure 2. A comparison of the three buffer ex-change methods evaluated processing times (A), recovery (B), and pH accuracy (C). Results demonstrate that FormPrep’s automated buffer exchange process is comparable to both dialysis and centrifugal filters.

Protein Stability Study

Processing time, protein recovery, and pH accuracy are important aspects of evaluating buffer exchange, but it is also crucial to study the impact of each buffer exchange process on protein stability. Size exclusion chromatography (SEC) was used to determine monomer, aggregate, and low molecular weight species of mAb α, mAb β, mAb γ and an enzyme. mAb β will be used as a representative sample here. Data and results for the other molecules can be found at freeslate.com.

The SEC results for mAb β show that FormPrep does not negatively impact the long-term stability of the processed biotherapeutics when compared to dialysis and centrifugal filters (Figure 3).

Accelerated stability results (Figure 3B, 3C, and 3D) indicated no significant differences in mAb β monomer, aggregate, and low molecular weight species at any time point regardless of buffer exchange method. These results were consistent for the other molecules tested by SEC, indicating that all three buffer exchange methods did not change the stability of these proteins, even after incubation at 40°C for six weeks.

In addition to quantifying soluble species by SEC, dynamic light scattering (DLS) and micro-flow imaging (MFI) were also used to further examine soluble species and sub-visible particles, respectively. DLS was used to monitor the radii and polydispersity of the four proteins after buffer exchange by centrifugal filters, dialysis, and FormPrep. 50 µL of each sample was prepared in analytical triplicate for analysis at 25°C in a Wyatt DynaPro Plate Reader.

After buffer exchange by centrifugal filters, dialysis, and FormPrep, the monomer radii of the mAb β were measured as 2.9 nm, 2.8 nm, and 3.1 nm respectively, while the unprocessed control radius was found to be 3.1 nm. To count sub-visible particles, MFI data was collected on a MFI5100. Using mAb β as our representative protein, particle counts after 6 weeks at 40°C ranged from 10,000 to 13,000 total particles for all processes with no significant differences seen between the three buffer exchange methods. These results indicate that no significant change to mAb β stability occurred after buffer exchange, and this result is consistent for all molecules in this study.

Figure 3. mAb ß SEC results from the accelerated stability study including chromatogram comparisons at T0 (3A),and the accelerated stability profiles of monomer, aggregate, and low molecular weight species (3B, 3C, 3D). All SEC data was collected on an Aglient 1200 with a Tosoh TSKgel UltraSW Aggregate column, 2x PBS mobile phase, and 50 µg injections.


This study has shown that an automated buffer exchange workflow, FormPrep, provides comparable results to dialysis and centrifugal UF/DF filters. After comparing results from several biopharmaceuticals molecules, FormPrep clearly meets or exceeds the same standards of protein recovery, pH accuracy, and protein stability at T0 and after storage for six weeks at 40°C as competing dialysis and centrifugal UF/DF methods. Typically formulation preparation, buffer exchange, and sample concentration require two to three days of a scientist’s time.

Automating this entire process with FormPrep liberates scientists and increases the number of formulations and protein concentrations they can evaluate within project timelines. FormPrep’s automated workflow can increase the repeatability and reproducibility of formulation preparation across projects, departments, and campaigns.    

Susan Darling ([email protected]) is director of marketing, Russell Burge, Ph.D., works as an applications scientist, Gregory Manley Ph.D., is a scientist, Sudheer Sami serves as an applications scientist, and Sowmya Tummala is a research associate at Freeslate.

Previous articleEpigenetics Research Reveals a Range of Clinical Possibilities
Next articleNIH Commits $260M to Centers for Common Disease Genomics