pplication Note comparison of automated and manual buffer exchange methods Introduction uffer preparation, exchange and sample concentration for a formulation screen can take 2 4 days of a scientist s time. While many labs have developed strategies to streamline formulation development, it s still relatively manual and requires significant resources which can limit the number of formulations evaluated. Unchained Labs has eliminated this bottleneck with a new, easy-to-use system that automates formulation preparation, buffer exchange and sample concentration. The (Figure 1) enables scientists to easily, rapidly and automatically generate up to 12 different protein formulations at the 1 8 ml scale with minimal hands-on time. The starts by automatically preparing buffers from stock solutions and, if necessary, titrating the buffers with acids or bases. Next, the buffer exchanges a protein into each buffer to make a final liquid protein formulation by using an innovative ultrafiltration/diafiltration (UF/DF) method that is well suited for buffer exchanging and concentrating biopharmaceuticals. fter buffer exchange is complete, the automatically concentrates the protein in each final formulation based on user input. Through the user-friendly software interface all technical staff can prepare protein formulations without any specialized training, resulting in an increase in formulation preparation throughput. Figure 1: The prepares 12 protein solutions automatically. utomated buffer exchange of therapeutic proteins using the The dispenses protein into single use filtration cups, measures the volume in each cup and then uses pressure to perform ultrafiltration (Figure 2). Filtration cups can be gently mixed during ultrafiltration to avoid clogging the membrane. fter filtration, the volume of each cup 1
STEP 1 dd protein STEP 2 Measure Volume STEP 3 pressure-based Filtration STEP 4 Measure Volume STEP 5 dd new uffer STEP 6 concentrate (optional) protein (mb, etc) uffer 1 uffer 2 Figure 2: Protein is added to each filtration cup and then volumes are measured. Pressure is applied to the filtration cup along with orbital mixing (optional) to concentrate the sample. The automatically measures solution volumes in each filtration cup, calculates the volume that was removed, refills with new formulation buffer and calculates percent exchange. The system repeats filtration, volume measurement and refill cycles until each sample reaches greater than 99% buffer exchange. is measured and the amount of buffer removed is calculated. The system refills each cup with the desired formulation buffer and calculates the percent (%) exchange. The repeats this process until each protein formulation is at least 99% buffer exchanged. Comparison of automated and traditional buffer exchange methods key question about automated processes is their comparability to traditional methods. For protein formulation, the buffer exchange process must provide excellent material recovery and not impact short term or long term protein stability. To answer these questions, we compared the s automated buffer exchange process to traditional methods by testing multiple proteins immediately after buffer exchange and on accelerated stability. The overall study design is shown in Figure 3. Four confidential therapeutic proteins were buffer exchanged using the, centrifugal UF devices and dialysis cassettes. Three of the proteins were monoclonal antibodies (mbs α, β and γ) and one was a therapeutic enzyme. The, centrifugal devices and dialysis were used to buffer exchange each protein into a target formulation by replacing 99% of the original buffer. The protein concentrations for the formulations were: mb α at 1 mg/ml, mb β at 5 mg/ml, mb γ at 5 mg/ml and a therapeutic enzyme at 2 mg/ml. fter buffer exchange, Unchained Labs' scientists determined protein recovery, the ph of each formulation and compared the stability of the proteins after incubation at 4 C over the course of six weeks. Protein formulation stability was monitored by size exclusion chromatography (SEC), dynamic light scattering (DLS), ph and micro-flow imaging (MFI). Immediately after buffer exchange (T ), each protein formulation was tested by SEC, DLS, ph and MFI. n aliquot from each sample (the, and dialysis) was removed, vialed in duplicate and then stored at 4 C. Each vialed aliquot was analyzed by SEC and DLS at, 2, 4 and 6 weeks, while MFI measurements were taken at and 4 weeks. For this study, the was used only to buffer exchange proteins. uffer preparation, protein concentration and additional capabilities of the are covered in other application notes available from Unchained Labs. Recovery and ph accuracy and centrifugal UF/DF filters, such as, are widely used methods for buffer exchange in formulation labs. is gentle on proteins but is slow and does not allow for concentration. filters are not as gentle on the protein but are faster and have the additional capability of concentration. oth methods provide high protein recovery and achieve satisfactory exchange efficiency but require significant hands-on time. 2
mb α mb β mb γ enzyme uffer exchange buffer exchange Recovery ph accuracy nalyze t, incubate @ 4 c, nalyze @ 2, 4,6 weeks Sec DlS MFi (t and 4 weeks) Recovery ph ccuracy accuracy nalyze t, incubate @ 4 c, nalyze @ 2, 4,6 weeks Sec Dl DlS MFi (t and 4 weeks) Recovery ph ccuracy accuracy nalyze t, incubate @ 4 c, nalyze @ 2, 4,6 weeks Sec Dl DlS MFi (t and 4 weeks) Figure 3: Stability study design for comparison of buffer exchange methods. Recovery Protein recovery was calculated by comparing the mass of protein before and after buffer exchange (Table 1). Protein recovery after buffer exchange by the ranged from 95.8 99.6% of expected mass. Mass recovery for the same protein formulations buffer exchanged with centrifugal concentrators and dialysis ranged from 91.1 94.3% and 91.6 96.4%, respectively. Protein recoveries for proteins buffer exchanged by the are highly comparable to and dialysis recoveries and in several cases, the has been shown to outperform both. ph accuracy The ph of each protein formulation postprocessing was compared to the target ph of the formulation buffer for all three buffer exchange techniques (Table 2). For all buffer exchange methods evaluated, ΔpH values of final protein formulations were within.9 ph units of targets. In contrast, when the two higher concentration protein formulations (mb β at 5 mg/ml and enzyme at 2 mg/ml) were buffer exchanged using dialysis cassettes, the ph values of the final formulation were.28.37 ph units from target. These results show that the is capable of meeting strict ph requirements for protein formulations after buffer exchange. The produced superior results in achieving ph targets for the final solution compared to dialysis for formulations with 2 mg/ml protein. No significant changes in ph values were observed for any of the buffer exchanged proteins during the stability study (data available upon request). ccelerated stability study results Size-Exclusion Chromatography (SEC) results To investigate the stability of each protein formulation after buffer exchange, mb α, mb β, mb γ and a therapeutic enzyme were incubated at 4 C for six weeks. SEC was used to determine the content of monomer, aggregates and low molecular weight species (fragments) for each protein. Monomer, aggregate and fragment content 3
Protein Concentration Recovery mb α 1 mg/ml 96.1% 94.3% 91.6% mb β 5 mg/ml 99.6% 92.3% 96.4% mb γ 5 mg/ml 95.8% 91.4% 96.% Enzyme 2 mg/ml 96.1% 91.1% 93.1% Table 1: The provides excellent protein recovery and is comparable to and dialysis. Protein Concentration ph Target mb α 1 mg/ml 5.9 -.1.1 -.11 mb β 5 mg/ml 5.88.9.6.28 mb γ 5 mg/ml 6.81 -.5 -.3. Enzyme 2 mg/ml 6.2...37 Table 2: Summary of the ph results post-processing for the three mbs and one enzyme after buffer exchange by the, and dialysis. of mb α, mb β, mb γ and the enzyme are highly similar for all time points and buffer exchange methods (Figures 4-7). The high degree of similarity indicates that all buffer exchange methods do not significantly influence the stability of the monomer of mb α (Figure 4), mb β (Figure 5), mb γ (Figure 6) and the enzyme (Figure 7), even after incubation for six weeks at 4 C. ggregate or low molecular weight species content of each protein did not significantly differ across all methods of buffer exchange. Note that no fragments were detected in any of the enzyme samples. ccelerated stability Dynamic Light Scattering (DLS) results DLS experiments were performed to monitor the hydrodynamic monomer radii of four proteins after buffer exchange by the, and dialysis. Figures 8 11 present DLS results for each protein and buffer exchange method Unchained Labs tested at T and after six weeks at 4 C. DLS results indicate that no significant changes in monomer radii occur after buffer exchange any of the proteins or after six weeks of incubation at 4 C. ccelerated stability Micro-Flow Imaging (MFI) results MFI measurements allowed for the determination of particle counts for each of the therapeutic protein formulations at T and after four weeks of incubation at 4 C. MFI results of particle concentrations (number of particles/ml of sample) for all four proteins that were buffer exchanged using the, and dialysis are shown in Figures 12 15. n unprocessed control that was stored at 4 C is included for comparison. These results clearly indicate that protein solutions that were buffer exchanged using the had comparable numbers of particles (2 >1 µm) to formulations that were buffer exchanged by, dialysis or the unprocessed control. 4
1 99 Monomer content of mb α 1 99 Monomer content of mb β Main peak percent area 98 97 96 95 94 93 92 91 Main peak percent area 98 97 96 95 94 93 92 91 9 2 4 6 Weeks at 4 C 9 2 4 6 Weeks at 4 C High molecular weight peak percent area 3. 2.5 2. 1.5 1..5 ggregate content of mb α 2 4 6 Weeks at 4 C High molecular weight peak percent area 5 4 3 2 1 ggregate content of mb β 2 4 6 Weeks at 4 C C Low molecular weight peak percent area 1 9 8 7 6 5 4 3 2 1 Fragment content of mb α 2 4 6 Weeks at 4 C Low molecular weight peak percent area 3. 2.5 2. 1.5 1..5 Fragment content of mb β 2 4 6 Weeks at 4 C Figure 4: uffer exchange methods did not significantly influence monomer, aggregate and fragment contents of 1 mg/ml mb α as evidenced by SEC results. : Percent area of protein monomer. : Percent area of protein aggregates (high molecular weight species or HMW). C: Percent area of low molecular weight species (LMW or fragments). Figure 5: Each buffer exchange method did not significantly influence monomer, aggregate and fragment contents of 5 mg/ml mb β as evidenced by SEC results. : Percent area of protein monomer. : Percent area of protein aggregates (high molecular weight species or HMW). C: Percent area of low molecular weight species (LMW or fragments). 5
1 98 Monomer content of mb γ 1 99 Monomer content of enzyme Main peak percent area 96 94 92 9 88 86 84 82 Main peak percent area 98 97 96 95 94 93 92 91 8 2 4 6 Weeks at 4 C 9 2 4 6 Weeks at 4 C High molecular weight peak percent area 2. 1.5 1..5 ggregate content of mb γ 2 4 6 Weeks at 4 C High molecular weight peak percent area 4. 3.5 3. 2.5 2. 1.5 1..5 ggregate content of enzyme 2 4 6 Weeks at 4 C C Low molecular weight peak percent area 1 9 8 7 6 5 4 3 2 1 Fragment content of mb γ 2 4 6 Weeks at 4 C Figure 7: ll buffer exchange methods did not significantly influence the monomer and aggregate content of 2 mg/ml enzyme as evidenced by SEC results. : Percent area of protein monomer. : Percent area of protein aggregates (high molecular weight species or HMW). Figure 6: ll buffer exchange methods did not significantly influence monomer, aggregate and fragment contents of 5 mg/ml mb γ as evidenced by SEC results. : Percent area of protein monomer. : Percent area of protein aggregates (high molecular weight species or HMW). C: Percent area of low molecular weight species (fragments). 6
6 5 mb α T 6 5 mb α T 6 5 mb α T 4 3 2 4 3 2 4 3 2 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 6 5 mb α 6 weeks 6 5 mb α 6 weeks 6 5 mb α 6 weeks 4 4 4 3 2 3 2 3 2 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 Figure 8: DLS results for 1 mg/ml mb α buffer exchanged using the, devices and dialysis cassettes. : T. : Six weeks at 4 C. 8 mb β T 1 8 mb β T 8 mb β T 6 4 6 4 6 4 2 2 2.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 9 8 mb β 6 weeks 9 8 mb β 6 weeks 9 8 mb β 6 weeks 7 7 7 6 5 4 3 6 5 4 3 6 5 4 3 2 2 2 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 Figure 9: DLS results for 5 mg/ml mb β buffer exchanged using the, devices and dialysis cassettes. : T. : Six weeks at 4 C. 7
8 7 mb γ T 8 7 mb γ T 8 7 mb γ T 6 6 6 5 4 3 5 4 3 5 4 3 2 2 2 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 8 7 mb γ 6 weeks 8 7 mb γ 6 weeks 8 7 mb γ 6 weeks 6 6 6 5 4 3 5 4 3 5 4 3 2 2 2 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 Figure 1: DLS results for 5 mg/ml mb γ buffer exchanged using the, devices and dialysis cassettes. : T. : Six weeks at 4 C. 8 7 Enzyme T 8 7 Enzyme T 8 7 Enzyme T 6 6 6 5 4 3 5 4 3 5 4 3 2 2 2 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 8 7 Enzyme 6 weeks 8 7 Enzyme 6 weeks 8 7 Enzyme 6 weeks 6 6 6 5 4 3 5 4 3 5 4 3 2 2 2 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1.1.1 1 1 1 1 1 Figure 11: DLS results for 2 mg/ml enzyme buffer exchanged using the, devices and dialysis cassettes. : T. : Six weeks at 4 C. 8
5 mb α particle counts 5 mb γ particle counts total particle counts 45 4 35 3 25 2 15 control GRUnt total particle counts 45 4 35 3 25 2 15 control GRUnt 1 1 5 5 t 4 weeks t 4 weeks Stability time point Figure 12: MFI results for 1 mg/ml mb α buffer exchanged using the, devices and dialysis cassettes. Stability time point Figure 14: MFI results for 5 mg/ml mb γ buffer exchanged using the, devices and dialysis cassettes. 3 mb β particle counts 5 enzyme particle counts total particle counts 25 2 15 1 control GRUnt total particle counts 45 4 35 3 25 2 15 control GRUnt 5 1 5 t 4 weeks t 4 weeks Stability time point Figure 13: MFI results for 5 mg/ml mb β buffer exchanged using the, devices and dialysis cassettes. Stability time point Figure 15: MFI results for 2 mg/ml enzyme buffer exchanged using the, devices and dialysis cassettes. 9
Conclusion This study has shown that an automated buffer exchange system, the, provides comparable results to centrifugal UF/DF filters and dialysis. fter comparing results from several biotherapeutic molecules, the clearly meets or exceeds the same standards of protein recovery, ph accuracy and protein stability at T and after storage for six weeks at 4 C as compared to centrifugal UF/DF filter and dialysis methods. Typically, formulation preparation, buffer exchange and sample concentration require 2 4 days of a scientist s time. utomating this process with the increases the number of formulations and protein concentrations they can evaluate within project timelines. The ability to prepare 12 formulations at high concentrations at a small scale also allows scientists to evaluate high protein concentrations earlier in development to assess developability and manufacturability. The s automated workflow can increase the repeatability and reproducibility of formulation preparation across projects, departments and campaigns. Unchained Labs' : Requires significantly less hands-on time than current buffer exchange methods. Does not significantly alter protein stability. Is compatible with a wide variety of proteins. chieves 99% exchange for most proteins. Unchained Labs 694 Koll Center Pkwy, Suite 2 Pleasanton, C 94566 Phone: 1.925.587.98 Toll-free: 1.8.815.6384 Email: info@unchainedlabs.com 216 Unchained Labs. ll rights reserved. Unchained Labs is a registered trademark. ll other brands or product names mentioned are trademarks owned by their respective organizations. Rev 1