Bioseparations and Downstream Processing

Biotechnology Progress DOI 10.1002/btpr.2024

The Inactivation of Viruses Using Novel Protein A Wash Buffers AUTHOR NAMES Glen R. Bolton, Keith Selvitelli, Ionela Iliescu, Doug Cecchini AUTHOR ADDRESS Biogen, 14 Cambridge Center, Building 8, Floor 3, O18, Cambridge, MA 02142 AUTHOR EMAIL ADDRESS [email protected] ABSTRACT Low pH viral inactivation is typically performed in the eluate pool following the protein A capture step during the manufacturing of monoclonal antibodies and Fc-fusion proteins. However exposure to low pH has the potential to alter protein quality. To avoid these difficulties, novel wash buffers capable of inactivating viruses while antibodies or Fc-fusion proteins were bound to protein A or mixed mode resins were developed. By equilibrating the column in high salt buffer (2 M ammonium sulfate or 3 M sodium chloride) after loading, the hydrophobic interactions between antibodies and protein A ligands were increased enough to prevent elution at pH 3. The ammonium sulfate was also found to cause binding of an antibody to a mixed mode cation exchange and a mixed mode anion exchange resin at pH values that caused elution in conventional cation and anion exchange resins (pH 3.5 for Capto Adhere and pH 8.0 for Capto MMC), indicating that retention was due to enhanced hydrophobic interactions. The potential of the 2 M ammonium sulfate pH 3-3.5 buffer, a 1 M arginine buffer, and a buffer containing the detergent LDAO to inactivate XMuLV virus when used as protein A wash buffers with a 1 hour contact time were studied. The high salt and detergent containing wash buffers provided about five logs of removal, determined using PCR, and complete combined removal and inactivation (> 6 logs), determined by measuring infectivity. The novel protein A washes could provide more rapid, automated viral inactivation steps with lower pool conductivities. KEYWORDS monoclonal antibody, Fc-fusion, protein A, mixed mode, purification, viral inactivation, viral clearance INTRODUCTION AND BACKGROUND As of January 2014, there have been 33 therapeutic monoclonal antibodies approved for use in the US or Europe (http://www.antibodysociety.org/news/approved_mabs.php). Eight have at least a billion dollars in annual revenue and together they comprise at least This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/btpr.2024 © 2014 American Institute of Chemical Engineers Biotechnol Prog Received: Oct 09, 2014; Revised: Dec 02, 2014; Accepted: Dec 03, 2014 This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

20 billion dollars in annual revenue 1. In addition, there are currently approximately 300 antibodies in clinical development 2. The number of licensed antibodies is growing annually at a rate of about 11% 3 and the number of clinical antibody candidates has been growing at a rate of 18% since 2004 2. The standard process for the purification of a monoclonal antibody from harvested cell culture fluid (HCCF) typically involves at least two chromatography steps: protein A affinity and anion exchange chromatography. Protein A chromatography separates proteins from impurities based on a reversible interaction between the Fc portion of a molecule and a protein A ligand immobilized on a chromatography matrix. Impurities flow through the matrix while the product stays bound to the column. The column is subsequently washed with solvent, detergents or salts to achieve additional impurity removal 4. This product is eluted from the matrix by a reduction in pH. Protein A is used to capture antibodies after the cell removal and clarification steps for the majority of the licensed antibodies. During the manufacturing of protein therapeutics derived via mammalian cell culture systems, the capacity of the purification process to effectively remove and/or inactivate virus must be demonstrated before the material is considered suitable for use in humans. Viral clearance can be achieved by removal or inactivation of viruses during the purification process5. An analysis of the retroviral clearance provided by the protein A step in 20-25 years of Investigational New Drug and Biologic License Application submissions to the FDA for antibody therapeutics indicates that the protein A step provides between 2 and 6 logs of viral clearance 6. In about half the cases, complete viral clearance was achieved. However the analysis did not indicate if a PCR assay or an infectivity assay was used to measure virus titer. The PCR assay would only measure virus removal, where the infectivity assay would measure combined removal and inactivation. In addition, varying degrees of viral inactivation may occur during protein A elution, depending on the elution buffer pH and buffer concentration, elution mass and volume, the elution flow rate, and how rapidly the pool was neutralized after elution. A low pH viral inactivation step is typically performed in the eluate pool following the protein A capture step during the manufacturing of monoclonal antibodies. This step typically provides at least 4 logs of viral inactivation 6. However exposure to low pH has the potential to alter protein quality, which may be observed as product precipitation, aggregation, inactivation, and instability 7. The negative effects of low pH can be greater at large scale where it can be time consuming and difficult to manually acidify, mix, transfer, and then neutralize a large pool. The acidification and subsequent neutralization can increase pool conductivity, which can impact the subsequent ion exchange chromatography step 8. The protein A step represents a substantial portion of the raw material cost in mAb manufacturing. Various methods have been employed to increase the productivity and utilization of protein A columns, including the use of dual flow rate loading strategies 9 and flow-through recycling in an effort to fully saturate the resin 10. Semi-continuous

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 2 of 19

Page 3 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biotechnology Progress

multi-column protein A chromatography has been shown 10, 11 to be an effective way to increase resin utilization and increase step throughput. This allows protein A to be loaded to a higher binding capacity, reducing the resin volume required and the overall buffer consumption. When performing semi-continuous multi-column chromatography, it is challenging to perform the low pH inactivation step on the many elution pools that are generated 12. Because of this difficulty, and because of the potential effects of low pH on product quality, it would be ideal to inactivate viruses while Fc containing molecules remain bound to the protein A resin. On column viral inactivation has been demonstrated for rFIX bound to an IMAC resin and mAbs bound to CEX resins 13, 14. However Fc containing molecules typically elute from protein A resin at the low pH values required for viral inactivation. An investigation into the modes of interaction between protein A ligands and Fc containing molecules was carried out to determine the feasibility of preventing product elution while achieving viral inactivation. The protein A ligand binds the Fc region of an antibody through a combination of hydrophobic, ionic and hydrogen bond interaction, as shown in Figure 1. The crystal structure of half the Fc portion of an antibody is shown binding to domain B of a protein A molecule through hydrogen bonding and hydrophobic interactions 15. The ionic interactions that are present are not depicted in the figure for simplicity. In this study, Experiments were performed to determine if the hydrophobic interactions between protein A and antibodies or Fc-fusion proteins could be increased enough using mobile phase modifiers to prevent the desorption that typically occurs at acidic pH values. MATERIALS AND METHODS Antibodies The monoclonal antibody (mAb) used for these studies was a humanized IgG1 produced in recombinant Chinese hamster ovary (CHO) cells grown in serum free medium. The mAb was purified using a protein A chromatography column and an anion exchange chromatography column. Both glycine and citrate were used as the low pH protein A elution and wash buffers. The choice of glycine or citrate is not expected to influence the results obtained in this study. The Fc-fusion protein used for these studies was produced in HEK293 cells grown in serum free medium. Either HCCF or material purified using a protein A chromatography column and two anion exchange chromatography columns was used. An Åkta 10 Explorer chromatography system (Cat. No. 18-1300-00, GE/Healthcare, Piscataway, NJ) and 0.66 cm ID columns were used in all of the laboratory-scale chromatography experiments. Xenotropic murine leukemia virus (XMuLV) was selected as a model virus to evaluate the ability of the steps to inactivate and remove retrovirus or retrovirus-like particles.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

PCR and infectivity based assays were used to quantify the amount of virus in the load and elution fractions in the studies. Samples were assayed for RNA detection and quantification by qPCR using commercially available primers and detection technology. Samples were extracted using a commercial silica-based kit and then concentrated and assayed in triplicate. The assay used optimized primer concentrations, divalent cation concentrations, annealing temperatures and cycling parameters. An XMuLV RNA standard was used for quantification. The samples were not nuclease treated. Infectivity was measured by successively diluting samples and measuring XMuLV plaques in well plates with PG4 indicator cells. Each sample was measured one time and the 95% confidence limit of the assay was less than ± 0.12 log10. Prior to the studies, experiments were performed to determine the appropriate dilution in which the buffer components did not interfere with the virus assay. The log reduction value (LRV) represented the log ratio of virus particles in the column load to that in the column eluate. Fractions of the eluate were collected and analyzed for monomer and aggregate content using analytical size exclusion chromatography (TSKgel G3000 SWXL, Tosoh BioScience, 08541). MAbSelect SuRe Protein A, Capto Adhere, and Capto MMC media were obtained from GE Healthcare. Fractogel EMD TMAE Hi Cap adsorbent was obtained from EMD Chemicals Inc. (1.16881, Gibbstown, NJ). A 2.1 x 30 mm Perseptive Biosystems Protein G immunodetection column (Applied Biosystems, Inc., 2-1002-00) with a 280 nm UV detector on the outlet was used to measure the concentration of Fc-fusion protein in the HCCF. RESULTS AND DISCUSSION EXAMPLE 1: Determination of Minimum Ammonium Sulfate Required to Prevent Antibody Elution from Protein A An experiment was done to determine the minimum concentration of ammonium sulfate required to keep the antibody bound to the protein A resin at low pH values that would typically cause product elution. A 6.6 mL MabSelect SuRe column (19.4 cm) was loaded with 50 mL of HCCF with 2.2 mg/ml antibody. The column was washed with 100 mM bis-tris (2-[bisamino]-2--1,3propanediol) buffer with 2000 mM ammonium sulfate at pH 6.6. A low pH wash buffer with ammonium sulfate was then applied (100 mM sodium citrate, 2000 mM ammonium sulfate pH 3.0). The ammonium sulfate concentration was then reduced to zero over a 9 CV gradient. The flow rate for all steps was 250 cm/hr. As shown in Figure 2, the antibody began to desorb at about 1700 mM and was fully eluted at about 200 mM ammonium sulfate. This indicates that at least 1700 mM ammonium sulfate is required to keep the antibody bound to the resin at low pH.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 4 of 19

Page 5 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biotechnology Progress

Since at least 1700 mM ammonium sulfate was required to keep the antibody bound to the resin at low pH, the solubility of the antibody was measured as a function of pH, antibody and ammonium sulfate concentration. The antibody was found to be soluble in 1000 mM ammonium sulfate but not 1500 mM ammonium sulfate. The precipitation was observed with 1500 mM ammonium sulfate at pH values between 3.5 and 8.0 and with antibody concentrations between 10 and 30 mg/mL. Therefore, it was not clear if the prevention of antibody elution at pH 3.0 was due to enhanced interaction with the protein A ligand or lack of antibody solubility in the elution buffer. Subsequent experiments with an anion exchange resin and mixed mode cation and anion exchange resins indicated that the prevention of elution was due to a combination of both reduced solubility and strengthened resin interactions. EXAMPLE 2: Use of a Low pH, High Ammonium Sulfate Wash with a mAb Bound to Protein A The goal of this experiment was to determine if using an isocratic 2 M ammonium sulfate pH 3.0 wash buffer would prevent product elution, enable on-column viral inactivation, enhance host cell protein removal, and not impact product quality. A 6.6 mL MabSelect SuRe column (19.4 cm) was first equilibrated and then loaded to 25.8 g/L resin using 50 mL of antibody in HCCF. A flow rate of 250 cm/hr was used except during the regeneration step (50 cm/hr). A pH 3.0, 2 M ammonium sulfate, 100 mM glycine wash was used to keep the antibody bound to the resin at low pH. The low pH, high ammonium sulfate wash was bracketed by a neutral, high ammonium sulfate wash buffer (pH 7.0, 2 M ammonium sulfate) to ensure that high levels of ammonium sulfate were present as the pH was lowered to 3.0 and also when it was subsequently raised to 7.0. Without this, significant product elution occurred before and after the wash step. Excess wash buffer was used at each step to ensure adequate buffer exchange. The protein concentration, conductivity, and pH in each step are shown in Figure 3. The buffers used in each step are shown in Table 1. Table 1. Step sequence and buffers used for Example 2. Step Equilibration Load Chase Wash 1 Wash 2 Wash 3 Wash 4 Elution Regeneration

Buffer 75 mM sodium phosphate, 100 mM sodium chloride Filtered HCCF 75 mM sodium phosphate, 100 mM sodium chloride 100 mM bis-tris, 2 M ammonium sulfate 100 mM glycine, 2M ammonium sulfate 100 mM bis-tris, 2 M ammonium sulfate 100 mM bis-tris 100 mM glycine 0.3 N sodium hydroxide

pH 7.3

Volume (CV) 5

7.3

7.6 3

7.0 3.0 7.0 7.0 3.0

5 5 5 6.3 3.5 5

John Wiley & Sons This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Slight product loss (1.9%) was observed during the 2 M ammonium sulfate, pH 3.0 wash. The majority of the product (94.4%) was recovered in the elution buffer. The percentage of the peak that was monomeric antibody, 96.0%, was not altered by the high ammonium sulfate wash. Viral inactivation was not measured in this experiment. The high salt, low pH wash buffer did not increase removal of host cell protein (HCP) on the protein A step. The levels of HCP in the eluate were typically the same or higher than those achieved using a simple neutral pH, sodium chloride wash (not shown). It is likely that the high levels of ammonium sulfate in the low pH wash enhanced interactions between HCP and antibodies or HCP and protein A resin. It is also possible that the ammonium sulfate decreased the solubility of the HCP, reducing the clearance. An additional experiment at conditions similar to the previous one was performed but with sodium chloride substituted for ammonium sulfate. The yield loss was higher using a 2 M sodium chloride, pH 3.0 wash step (17.0%) compared to the 2 M ammonium sulfate pH 3.0 wash step (1.9%). Since ammonium sulfate is a stronger kosmotrope than sodium chloride, a lower concentration is required to prevent antibody elution. It is believed that using a higher concentration of sodium chloride (3 M) would prevent antibody elution and allow on-column viral inactivation. In this case, the choice of salt would be governed by the cost and practicality of using a specific salt. The effects of the high salt buffers on fouling and leaching of the protein A ligand were not measured in this study. The conditions are not expected to result in elevated ligand leaching however as the pH and salt levels are within the recommended ranges for the resin and have been previously tested16. EXAMPLE 3: Inactivation and Removal of Viruses Using a Low pH, High Ammonium Sulfate Wash With An Fc-fusion Protein Bound to Protein A The feasibility of performing viral inactivation while an Fc-fusion protein was bound to a protein A resin was evaluated. The Fc-fusion protein was bound to the protein A and then a high salt neutral pH wash was applied to the column. A low pH, high salt (pH 3.5, 2 M ammonium sulfate) was then applied to inactivate virus while product remained bound to the adsorbent. A 7.2 mL MabSelect SuRe column (21 cm) was first equilibrated (EQ) with 4 column volumes (CVs) of 10 mM sodium phosphate, 140 mM sodium chloride, pH 7.4. Subsequently, 256 ml of HCCF containing Fc-fusion protein was loaded (25.5 mg per mL of resin) onto the column. The loading was followed by seven wash steps as indicated in Table 2 below. Excess wash buffer was used at each step to ensure adequate buffer exchange. The product was subsequently eluted with 25 mM citrate, 150 mM sodium chloride, pH 3.4 and quickly neutralized. The flow rate was 300 cm/hr except during the low pH wash step, which was slowed to 100 cm/hr to allow for a one hour contact time. The product recovery in the eluate was 84%.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 6 of 19

Page 7 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biotechnology Progress

Table 2. Step sequence and buffers used for Example 3. Step Equilibration Load Wash 1 Wash 2 Wash 3 Wash 4 Wash 5 Wash 6 Wash 7 Elution Strip Regeneration

Buffer 10 mM sodium phosphate, 140 mM sodium chloride HCCF 10 mM sodium phosphate, 140 mM sodium chloride 10 mM sodium phosphate, 900 mM sodium chloride 100 mM bis-tris, 2 M ammonium sulfate 100 mM sodium citrate, 2 M ammonium sulfate 100 mM bis-tris, 2 M ammonium sulfate 10 mM sodium phosphate, 900 mM sodium chloride 10 mM sodium phosphate, 140 mM sodium chloride 25 mM citrate, 150 mM sodium chloride 10 mM sodium phosphate, 900 mM sodium chloride 0.1 N sodium hydroxide

pH 7.4

Volume (CV or ml) 4

7.4

256 ml 4.5

7.4

4

7.0 3.5 7.0 7.4

4 5 (One hour contact time) 4 4

7.4

3

3.4 7.4

3 5 3

Viral titer was measured by infectivity and PCR. The retroviral removal was calculated to be 4.96 log10 using PCR. The combined retroviral removal and inactivation measured by infectivity was >6.39 log10, indicating that the 1 hour low pH/high salt wash buffer can provide effective inactivation of viruses. EXAMPLE 4: Use of a Low pH, High Ammonium Sulfate Wash with a mAb Bound to a Conventional and a Mixed Mode Anion Exchanger The goal of this experiment was to determine if high levels of ammonium sulfate would enhance hydrophobic interactions and prevent the low pH elution of an antibody bound to a resin (Capto Adhere), which has hydrophobic, hydrogen bonding, and anion exchange binding properties, potentially enabling on-column viral inactivation. A 1.7 mL (5.0 cm) Capto Adhere column was first equilibrated and then loaded with 6 mL of 11.4 g/L purified antibody in pH 8.0 50 mM tris buffer. Subsequently a high salt pH 8.0 wash buffer and then a high salt pH 3.5 wash buffer were applied to the resin. Excess wash buffer was used at each step to ensure adequate buffer exchange. All steps were performed at a linear flow rate of 100 cm/hr. The protein concentration, conductivity, and pH in each step are shown in Figure 4. The buffers used in each step are shown in Table 3. Table 3. Step sequence and buffers used for Example 4. Step Equilibration

Buffer 50 mM tris

pH 8.0

Volume (CV or ml) 15

John Wiley & Sons This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Load Wash 1 Wash 2 Wash 3 Wash 4 Elution Regeneration

Purified antibody in 50 mM tris 50 mM tris, 2M ammonium sulfate 100 mM citrate 2M ammonium sulfate 50 mM tris, 2M ammonium sulfate 50 mM tris 100 mM citrate 1.0 N sodium hydroxide

Page 8 of 19

8.0 8.0 3.5 8.0 8.0 3.4

6 mL 15 15 15 15 15 15

As shown in Figure 4, some product loss occurred during the loading and minimal product loss was observed during the 2 M ammonium sulfate, pH 3.5 wash, which is surprising because at this pH both the product and the resin have a positive charge. It is likely that the product remains bound to the resin due to enhanced hydrophobic interactions in this high salt buffer. The majority of the product (79.0%) was recovered in the elution pool. The percentage of the peak that was monomeric antibody (98.4%) was not altered by the high ammonium sulfate wash. A similar experiment was performed using TMAE HiCap resin. This is an anion exchange resin that lacks the hydrophobic interaction properties of the Capto Adhere resin. Significant product loss occurred during the low pH washes and the overall recovery was 49.8%. The product that eluted during the low pH wash precipitated in the column and in the instrumentation. This experiment indicates that kosmotropic salts were not capable of preserving adsorption of the mAb to the anion exchange resin at low pH. EXAMPLE 5: Use of a High pH, High Ammonium Sulfate Wash with a mAb Bound to Capto MMC The goal of this experiment was to determine if high levels of ammonium sulfate would enhance hydrophobic interactions and prevent the high pH elution of an antibody bound to a mixed mode cation exchange resin (Capto MMC). This resin has hydrophobic, hydrogen bonding, and cation exchange binding properties. Typically pH 8.0 wash buffers cause elution of antibodies from Capto MMC resin. A 1.7 mL (5.1 cm) column with Capto MMC mixed mode resin was loaded with 6 mL of 12.4 mg/mL antibody in a pH 4.5, 50 mM citrate buffer. Subsequently a high salt pH 4.5 wash and then a high salt pH 8.0 wash were applied. Excess wash buffer was used at each step to ensure adequate buffer exchange. The protein concentration, conductivity, and pH in each step are shown in Figure 5. The buffers used in each step are shown in Table 4. All steps were performed at a linear flow rate of 100 cm/hr.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 9 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biotechnology Progress

Table 4. Step sequence and buffers used for Example 5. Step Equilibration Load Wash 1 Wash 2 Elution Regeneration

Buffer 50 mM citrate Purified antibody in 50 mM citrate 50 mM citrate, 2M ammonium sulfate 50mM tris, 2M ammonium sulfate 50 mM tris 0.1 N sodium hydroxide

pH 4.5 4.5 4.5 8.0 8.0

Volume (CV or ml) 15 6 mL 15 15 15 15

As shown in Figure 5, minimal product loss was observed during the 2 M ammonium sulfate, pH 8.0 wash, which is surprising because at this pH both the product and the resin have a negative charge and antibody elution would be expected to occur. The percent recovery in the eluate pool was 94.1%. It is likely that the product remains bound to the resin due to enhanced hydrophobic interactions in this high salt buffer. While an unexpected finding, it is not clear what utility keeping antibodies bound to Capto MMC resin at pH 8.0 provides. Viruses can be inactivated by high pH but 8.0 is not sufficient. It is possible that the high pH wash buffer has the potential to improve impurity clearance. However high pH can affect product quality and low pH viral inactivation is typically performed so the utility of this high pH wash buffer is not clear. EXAMPLE 6: Inactivation and Removal of Viruses Using an Arginine Wash With A Fc-fusion Protein Bound to Protein A The use of neutral or acidic arginine-containing solutions to inactivate viruses has been previously described17-19. The mechanism of virus inactivation in the presence of arginine has been hypothesized to be destabilization of the viral envelope. An experiment was performed to determine if using a 1 M arginine, pH 4.7 protein A wash would provide viral inactivation while product remained bound to the resin. A 7.0 mL (20.5 cm) MabSelect SuRe column was first equilibrated and then 157 ml of HCCF containing Fc-fusion protein was loaded onto the column. A modified wash containing arginine was then applied for one hour to inactivate virus. The flow rate was 300 cm/hr except during the arginine wash step, which was slowed to 100 cm/hr to allow for a one hour contact time. Table 5 summarizes the buffer solutions used in each step. Table 5. Step sequence and buffers used for Example 6. Step Equilibration Load Wash 1 Wash 2 Wash 3

Buffer 10 mM sodium phosphate, 140 mM sodium chloride HCCF 10 mM sodium phosphate, 140 mM sodium chloride 10 mM sodium phosphate, 900 mM sodium chloride 1 M arginine hydrochloric acid

pH 7.4

Volume (CV or ml) 4

7.4

157 ml 4.5

7.4

4

4.7

5

John Wiley & Sons This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Wash 4 Elution Strip Regeneration

10 mM sodium phosphate, 140 mM sodium chloride 25 mM citrate, 150 mM sodium chloride 10 mM sodium phosphate, 900 mM sodium chloride 0.1 N sodium hydroxide

Page 10 of 19

7.4

3

3.4 7.4

3 5 3

The product recovery in the eluate was 83%. This was determined from the mass in the load, measured using protein G, and the mass in the pool, measured using absorbance at 280 nm. The viral removal as measured by PCR was >5.54 log10, and the combined removal and inactivation measured by infectivity was >6.39 log10. The viral inactivation was similar to that obtained in Example 3, indicating that either 1 M arginine pH 4.7 or 2 M ammonium sulfate pH 3.5 can be used for on-column viral inactivation, though the arginine removed the majority of virus before inactivation. EXAMPLE 7: Inactivation and Removal of Viruses Using a Detergent Wash With A Fc-fusion Protein Bound to Protein A Triton X-100 has often been used for inactivation of enveloped viruses. The use of a more environmentally friendly detergent, N,N- dimethyldodecylamine N-oxide (LDAO), for viral inactivation has recently been described 20. An experiment was performed to determine if using LDAO as a protein A wash buffer would provide viral inactivation and removal while product remained bound to the resin. Based on previous studies, an LDAO concentration four fold higher than the critical micelle concentration was used 20. A 7.0 mL (20.5) MabSelect SuRe column was equilibrated and then loaded with 256 ml of HCCF containing Fc-fusion protein. A modified wash containing LDAO was then applied for one hour to inactivate virus. The flow rate was 300 cm/hr except during the LDAO wash step which was slowed to 100 cm/hr to allow for a one hour contact time. Table 6 below summarizes the buffer solutions used in each step. Table 6. Step sequence and buffers used for Example 7. Step

Buffer

pH

Volume (CV or ml)

Equilibration

10 mM sodium phosphate, 140 mM sodium chloride

7.4

4

Load

HCCF

Wash 1

10 mM sodium phosphate, 140 mM sodium chloride

7.4

4.5

Wash 2

10 mM sodium phosphate, 900 mM sodium chloride

7.4

4

256 ml

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 11 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biotechnology Progress

Wash 3

10 mM sodium phosphate, 140 mM sodium chloride

3

Wash 4

10 mM sodium phosphate, 140 mM sodium chloride, 0.18% w/w LDAO

7.4

5

Wash 5

10 mM sodium phosphate, 140 mM sodium chloride

7.4

3

Elution

25 mM citrate, 150 mM sodium chloride

3.4

3

Strip

10 mM sodium phosphate, 900 mM sodium chloride

7.4

5

Regeneration

0.1 N sodium hydroxide

3

The product recovery in the eluate was 85%. This was determined from the mass in the load, measured using protein G, and the mass in the elution pool, measured using absorbance at 280 nm. The viral removal measured by PCR was 5.11 log10, and the combined removal and inactivation measured by infectivity was > 6.40 log10. This indicates that detergents such as LDAO have the potential for use for on-column viral inactivation and removal. SUMMARY AND CONCLUSIONS In this study the hydrophobic interactions between antibodies and protein A ligands were increased enough to prevent elution at pH 3 by adding 2 M ammonium sulfate. The ammonium sulfate was also found to cause binding of an antibody to a mixed mode cation exchanger and a mixed mode anion exchange at pH values that caused elution in conventional cation and anion exchange resins (pH 3.5 for Capto Adhere and pH 8.0 for Capto MMC). This suggests that retention was due to enhanced hydrophobic interactions and not due to a lack of solubility. The 2 M ammonium sulfate pH 3-3.5 buffer, a 1 M arginine buffer, and a buffer containing the detergent LDAO inactivated viruses when used as protein A wash buffers with a 1 hour contact time. The high salt and detergent containing wash buffers provided about five logs of removal, determined using PCR, and complete combined removal and inactivation (> 6 logs), determined by measuring infectivity. The arginine provided complete removal of XMuLV, as determined using PCR. When performing semi-continuous multi-column chromatography, it is challenging to perform the low pH inactivation step on the many elution pools that are generated. The novel protein A washes could provide more rapid, automated viral inactivation steps with lower pool conductivities. However the additional buffer washes and the on-column inactivation will add processing time and complexity. In addition, the high ionic strength buffers would have to be evaluated for compatibility with process equipment and materials of contact. ACKNOWLEDGMENTS The authors thank Hiro Aono, Sanchayita Ghose, Rich Macniven, John Pieracci, Matt Westoby and Jennifer Zhang

John Wiley & Sons This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

REFERENCES 1. Yamane-Ohnuki, N.; Satoh, M., Production of therapeutic antibodies with controlled fucosylation. mAbs 2009, 1, (3), 230-236. 2. PhRMA, P. R. a. M. o. A., Biotechnology Medicines in Development. In Pharmaceutical Research and Manufacturers of America (PhRMA): 2011. 3. Kelley, B., Industrialization of mAb production technology: The bioprocessing industry at a crossroads. mAbs 2009, 1, (5), 443-452. 4. Shukla, A. A.; Hinckley, P., Host cell protein clearance during protein A chromatography: development of an improved column wash step. Biotechnol Prog 2008, 24, (5), 1115-21. 5. Miesegaes, G. R.; Lute, S. C.; Read, E. K.; Brorson, K. A., Viral clearance by flow-through mode ion exchange columns and membrane adsorbers. Biotechnol Prog 2014, 30, (1), 124-31. 6. Miesegaes, G.; Lute, S.; Brorson, K., Analysis of viral clearance unit operations for monoclonal antibodies. Biotechnology and Bioengineering 2010, 106, (2), 238-246. 7. Bywater, R.; Eriksson, G. B.; Ottosson, T., Desorption of immunoglobulins from Protein A-Sepharose CL-4B under mild conditions. J Immunol Methods 1983, 64, (1-2), 1-6. 8. Curtis, S.; Lee, K.; Blank, G. S.; Brorson, K.; Xu, Y., Generic/matrix evaluation of SV40 clearance by anion exchange chromatography in flow-through mode. Biotechnology and Bioengineering 2003, 84, (2), 179-186. 9. Ghose, S.; Nagrath, D.; Hubbard, B.; Brooks, C.; Cramer, S. M., Use and Optimization of a Dual-Flowrate Loading Strategy To Maximize Throughput in ProteinA Affinity Chromatography. Biotechnology Progress 2004, 20, (3), 830-840. 10. Mahajan, E.; George, A.; Wolk, B., Improving affinity chromatography resin efficiency using semi-continuous chromatography. Journal of Chromatography A 2012, 1227, (2), 154-162. 11. Pollock, J.; Bolton, G.; Coffman, J.; Ho, S. V.; Bracewell, D. G.; Farid, S. S., Optimising the design and operation of semi-continuous affinity chromatography for clinical and commercial manufacture. Journal of Chromatography A 2013. 12. Ransohoff, T. C.; Bisschops, M. A. T., Continuous processing methods for biological products. In Patent Application WO2012078677 A2: 2013. 13. Roberts, P. L.; Walker, C. P.; Feldman, P. A., Removal and Inactivation of Enveloped and Non-Enveloped Viruses during the Purification of a High-Purity Factor IX by Metal Chelate Affinity Chromatography. Vox Sanguinis 1994, 67, 69-71. 14. Soice, N.; Hubbard, J.; Zhang, Y.; Hamzik, J., Methods for Purifying a Target Protien From One of More Impurities in a Sample. In United States Patent Applicaiton US 2011/0065901 A1: 2011. 15. Deisenhofer, J., Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8-.ANG. resolution. Biochemistry 1981, 20, (9), 2361-2370. 16. Grönberg, A.; Eriksson, M.; Ersoy, M.; Johansson, H. J., A tool for increasing the lifetime of chromatography resins. mAbs 2011, 3, (2), 192-202. 17. Kozloff, L. M.; Lute, M.; Crosby, L. K.; Wong, R.; Stern, B., Critical arginine residue for maintaining the bacteriophage tail structure. Journal of virology 1969, 3, (2), 217-27.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 12 of 19

Page 13 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biotechnology Progress

18. Yamasaki, H.; Tsujimoto, K.; Koyama, A. H.; Ejima, D.; Arakawa, T., Arginine facilitates inactivation of enveloped viruses. Journal of Pharmaceutical Sciences 2008, 97, (8), 3067-3073. 19. McCue, J. T.; Selvitelli, K.; Cecchini, D.; Brown, R., Enveloped virus inactivation using neutral arginine solutions and applications in therapeutic protein purification processes. Biotechnology Progress 2014, 30, (1), 108-112. 20. Conley, L. E.; Tao, Y., Methods and compositions for inactivating enveloped viruses. In Patent Application WO2014025771 A2: 2014.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

LIST OF FIGURES Figure 1: The crystal structure of half the Fc portion of an antibody binding to domain D of a protein A molecule 15. The Fc CH2 and CH3 domains are shown in red. The domain B of the protein A (shown in blue) binds the Fc at the elbow of the CH2 and CH3 domain. The four hydrogen bonds are shown as dotted red lines. In the center of the image, the phenylalanine 132 of protein A can be seen closely interacting with the isoleucine 253 of the Fc through hydrophobic interactions. The ionic interactions that are present are not depicted in the figure for simplicity. Figure 2: The concentration of ammonium sulfate was reduced from 2 M to zero over a 9 CV gradient at pH 3.0 to determine the minimum concentration of ammonium sulfate required to keep the antibody bound to the protein A resin at low pH values. At least 1700 mM ammonium sulfate is required to keep the antibody bound to the resin at low pH. Figure 3: Protein concentration, conductivity and pH versus protein A column volumes using a pH 3.0, 2 M ammonium sulfate, 100 mM glycine wash. The high level of ammonium sulfate prevented the low pH elution of the antibody, potentially enabling oncolumn viral inactivation. Figure 4: Protein concentration, conductivity and pH versus column volumes using a pH 3.5, 2 M ammonium sulfate, 100 mM citrate wash to keep the antibody bound to the mixed mode anion exchange resin (Capto Adhere) resin at low pH. Figure 5. Protein concentration, conductivity and pH versus column volumes using a pH 8.0, 2 M ammonium sulfate, 50 mM phosphate wash to keep the antibody bound to the mixed mode cation exchange resin (Capto MMC) resin at high pH.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 14 of 19

Page 15 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biotechnology Progress

Figure 1: The crystal structure of half the Fc portion of an antibody binding to domain D of a protein A molecule (Deisenhofer 1981). The Fc CH2 and CH3 domains are shown in red. The domain B of the protein A (shown in blue) binds the Fc at the elbow of the CH2 and CH3 domain. The four hydrogen bonds are shown as dotted red lines. In the center of the image, the phenylalanine 132 of protein A can be seen closely interacting with the isoleucine 253 of the Fc through hydrophobic interactions. The ionic interactions that are present are not depicted in the figure for simplicity.

John Wiley & Sons This article is protected by copyright. All rights reserved.

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 16 of 19

2000 Ammonium Sulfate (mM)

Absorbance 280 nm (mAu)

1000

0 27

37 CV John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 17 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Biotechnology Progress

pH Absorbance 280 nm (mAU) Conductivity (mS/cm)

0

Column Volumes

John Wiley & Sons This article is protected by copyright. All rights reserved.

40

Biotechnology Progress

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 18 of 19

pH Absorbance 280 nm (mAU) Conductivity (mS/cm)

0

100 Column Volumes

John Wiley & Sons This article is protected by copyright. All rights reserved.

Page 19 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Biotechnology Progress

pH Absorbance 280 nm (mAU) Conductivity (mS/cm)

0

75 Column Volumes

John Wiley & Sons This article is protected by copyright. All rights reserved.

Inactivation of viruses using novel protein A wash buffers.

Low pH viral inactivation is typically performed in the eluate pool following the protein A capture step during the manufacturing of monoclonal antibo...
612KB Sizes 4 Downloads 5 Views