Introduction
Studies presented at the 2011 Viral Clearance Symposium served to further understand the critical process parameters (CPPs) involved in effective low-pH viral inactivation, and to define worst-case conditions regarding pH, temperature, hold duration, and ionic strength (1). One of the outcomes of this work was the development of a low-pH inactivation protocol that could guide firms and contract testing organizations in the design and performance of low-pH inactivation studies (2). Generally, log reduction values (LRVs) have been shown to increase (i.e., greater virus inactivation) as the temperature, hold duration, and protein concentration is increased, while the pH is decreased. A new study presented at the 2013 symposium discussed the effect of another parameter, buffer system composition, on LRVs after low-pH treatment.
In addition to low-pH inactivation, treatment with detergents is a very robust and effective method for inactivation of enveloped viruses such as Xenotropic murine leukemia virus (XMuLV). From the 2009 and 2011 symposia, the CPPs for viral inactivation were studied (i.e., temperature, time, and detergent concentration); however, it was noted that further investigations were required to identify and confirm other potential platform-specific CPPs (3). Also, additional experiments would be useful in supporting the use of bracketed operating conditions for CPPs across different product lines and using multiple viruses. One of the presentations at the 2013 symposium discussed the effect that changes in process conditions have on the kinetics of inactivation by Triton X-100.
Some of the additional challenges faced by industry in the use of detergents during biopharmaceutical manufacturing are the need to reduce and quantify residual levels of detergent in the final product, and the proper disposal of detergents according to local bylaws. Although presentations at the 2011 symposium showed promising results for several new generation detergents, it was agreed that work was still needed in collaboration with detergent manufacturers to establish the safety of their ingredients. Two of the presentations at the 2013 symposium discussed the use of eco-friendly and cost-effective detergents for viral inactivation as alternatives to the non-ionic detergent Triton X-100.
Virus Inactivation At Low Ph: Impact of Buffers, Ph Value, Protein Concentration, and Temperature On the Inactivation of Retroviruses (Franz Nothelfer, Hans-Georg Bischof, Helmut Winter, Joey Studts, Boehringer Ingelheim Pharma Gmbh & Co. Kg; Alice Marinaci and Horst Ruppach, Charles River Biopharmaceutical Services)
Recombinant proteins for pharmaceutical applications are predominantly produced using recombinant suspension cell cultures. To ensure product safety regarding absence of viruses, the respective cell cultures and raw materials are tested. In addition, at least two different orthogonal steps in the purification process are selected to remove and/or inactivate potential viruses. Small-scale virus clearance studies are performed to determine the virus reduction capability of the process and thereby demonstrate the safety of the biological product. To validate the potential of the virus removal of downstream processes, different model viruses are established. For potential retroviruses, the model virus XMuLV is frequently used to evaluate the inactivation at low pH. Case studies will show the impact of different buffer components like sodium acetate, sodium citrate, and glycine/phosphate on the inactivation of XMuLV during acid treatment. Other important parameters like pH value, protein concentration, and temperature are presented. When low-pH inactivation study results derived from various labs using various experimental conditions were plotted, somewhat unexpected virus inactivation results were observed (Figures 1⇓–3).
Viral inactivation of XMuLV in phosphate buffer: illustration of a chromatography elution pool that is forward-processed to the viral filtration step over two cycles. EQ = Equilibration; Regen.= Regeneration.
Viral inactivation of XMuLV in citrate buffer: proposed bracketed viral clearance strategy for viral filter cycling.
Viral inactivation of XMuLV in acetate buffer.
Potential reasons for variation in XMuLV inactivation were assumed to be the following: different contract labs, different strains of XMuLV, different buffer salts, and different protein concentrations. To further explore the potential effect that buffer salts might have on XMuLV inactivation, three different buffer systems were tested at pH 3.3, pH 3.6, and pH 3.9 at different time points. In addition, the effects of protein concentration and inactivation temperature were evaluated.
With the citrate buffer system (Figure 4), at the 60 min time point the inactivation of XMuLV at pH 3.3 was effective at a temperature of 20 °C and 30 °C at low and high protein concentrations. The inactivation of XMuLV at pH 3.3 at a temperature of 15 °C was moderate. No inactivation or only marginal inactivation of XMuLV at pH 3.6 and pH 3.9 at 15 °C and 20 °C at low and high protein concentration was shown at the 60 min time point.
Effect of pH, temperature and protein concentration on XMuLV inactivation in citrate buffer.
With the acetate buffer system (Figure 5), at the 30 min and the 60 min time points, the inactivation of XMuLV at pH 3.3 to pH 3.9 was very effective. The protein concentration and the temperature had no impact on the inactivation effectiveness.
Effect of pH and protein concentration on XMuLV inactivation in glycine-phosphate buffer.
With the glycine/phosphate buffer system (Figure 6), at the 60 min time point the inactivation of XMuLV at pH 3.3 was effective. At pH 3.6 a moderate inactivation was shown. No clearance was achieved at pH 3.9. The protein concentration had no impact on the inactivation effectiveness.
Effect of pH, temperature and protein concentration on XMuLV inactivation in acetate buffer.
In conclusion, the sodium acetate buffer system is considered very robust for conducting viral inactivation studies because it inactivates XMuLV completely at all tested pH values, temperatures, and protein concentrations. In comparison, the inactivation of XMuLV in the sodium citrate buffer system is affected by the pH value, protein concentration, and temperature. The inactivation of XMuLV in the glycine/phosphate buffer system is affected by the pH value but not by the protein concentration (the impact of temperature was not tested). The sodium citrate buffer and glycine/phosphate buffer systems are, therefore, robust at a pH value of 3.3; however, inactivation at higher pH values is harder to predict.
Method of Inactivating Enveloped Viruses Using Eco-friendly, Cost Effective, Zwitterionic Detergents (Lynn Conley, Biogen Idec)
Solvent/detergent viral inactivation methods that use an organic solvent such as tri-n-butyl phosphate (TnBP) and non-ionic detergents such as Polysorbate 80 or Triton X-100, or Triton X-100 alone without detergent, have been shown to be effective at inactivating enveloped viruses. These detergents may pose economic and wastewater concerns in some countries because of the ecotoxic nature of the detergent and may require costly disposal methods.
Thirty non-ionic and zwitterionic detergents that were known to be non-denaturing when solubilizing proteins from cell membranes were considered as an alternative means of achieving viral inactivation in harvested cell culture fluid (HCCF). These detergents were assessed based on cost (relative to Triton X-100) and their ecotoxicity according to the Danish ABC wastewater classification system. Danish wastewater class A chemicals, such as Triton X-100, must be eliminated or substituted due to potential irreversible human harm and/or high aquatic toxicity. Class B substances can be discharged but the concentrations must be within certain limits in order to make sure that water quality criteria are observed. Class C includes chemicals that can be discharged without concentration limits. Four zwitterionic detergents, sulfobetadine (SB) 3-12, SB-14, SB 3-16, and lauryldimethylamine N-oxide (LDAO) all met the desired criteria for wastewater ecotoxicity (either class B or C) and cost of the detergent. A feasibility study was performed to evaluate the ability of all four detergents to inactivate XMULV in HCCF that contained a recombinant Fc (immunoglobulin crystallizable fragment) fusion protein. Each detergent was added to a final concentration that corresponded to 10× the critical micelle concentration (CMC) in water. After detergent addition, the samples were held at ambient temperature (∼20 °C) for 120 min. All four detergents showed ≥4 LRV for XMULV as shown in Table I. In addition, none of the detergents affected the biological activity of the molecule with the binding activity being >90% after detergent treatment relative to an untreated control.
Inactivation of XMULV and Binding Activity of an Fc Fusion Protein in HCCF after Incubation with Zwitterionic Detergents for 120 min at ∼20 °C
Based on the feasibility study results, viral inactivation with LDAO was selected for further optimization and use in the clinical manufacturing processes. LDAO is moderately toxic to aquatic organisms, but rapidly degrades in water to non-hazardous components (4). LDAO is a class C substance in the Danish wastewater classification system and it can be discharged into wastewater without concentration limits or monitoring. In contrast, the SB detergents are considered class B reagents and would require constant monitoring and concentration limits for discharge into wastewater.
The next study evaluated viral inactivation kinetics as a function of LDAO concentrations ranging from 0.012% to 0.138% (w/v) in HCCF containing a monoclonal antibody (MAb). An incubation temperature of 5 °C was selected as a worst-case condition for viral inactivation kinetics. Effective viral inactivation requires that detergent concentrations be at or above the CMC, which in turn can be affected by the pH, ionic strength, and temperature of the solution. For example, the CMC value for LDAO at pH 7 in 150 mM NaCl, which is similar to the pH and ionic strength of HCCF, was reported to be approximately 0.023% at ambient temperature (5). As shown in Table II, rapid inactivation of virus (≥2.8 logs) occurred within 15 min with LDAO concentrations at 0.023% or above, and XMULV was below the limit of detection at both 15 and 120 min. In contrast, at the lower LDAO concentration tested (i.e., 0.012%), the rate of inactivation was slower and there was still virus detected after 120 min of incubation.
Kinetics of XMULV Inactivation in HCCF at 5 °C as a Function of LDAO Concentration
The LDAO concentration of 0.138% (w/v) was chosen for viral inactivation of enveloped viruses in HCCF for our clinical manufacturing processes. This detergent concentration is sufficiently above the CMC to account for varying cell culture conditions and will ensure robust inactivation of enveloped viruses. To date, LDAO has been used for viral inactivation in HCCF for three MAbs and one recombinant Fc fusion protein. The results of the viral inactivation studies for XMULV and Suid herpes virus (also known as Pseudorabies virus) are summarized in Table III. The kinetics of XMULV virus inactivation at 5 °C was rapid, and virus levels fell below the limit of detection within 5 min for all molecules. The Suid herpesvirus kinetics of inactivation was also rapid, with virus levels falling below the limit of detection within 5 min. In one case, a low but detectable amount of Suid herpesvirus was found at the 15 min time point, although a LRV of 4.8 was achieved. LDAO was shown to be effectively removed by the subsequent protein A capture step, and the levels of residual LDAO in the protein A column elution pools were less than 2 ng/mL (the limit of detection). There was no detectable impact on quality attributes for all four of the molecules tested after 120 min of inactivation and subsequently in the drug substance for all of the clinical batches tested.
Summary of XMuLV and Suid Herpes Viral Inactivation Kinetics with LDAO for 3 MAbs and 1 Fc Fusion Protein at 5 °C
In summary, LDAO was been shown to be a cost-effective alternative to Triton X-100 for the inactivation of lipid enveloped viruses during the production of MAbs and recombinant FC fusion proteins. The detergent does not alter product quality and is ecofriendly, eliminating wastewater disposal constraints.
Investigation of Alternative and Eco-friendly Detergents for Viral Inactivation (S. Fisher and L. Norling, Genentech)
Initial Detergent Screen
Genentech has established robust virus inactivation conditions in HCCF using Triton X-100. However, due to disposal restrictions in some jurisdictions, alternative detergents were evaluated. Criteria in the selection of a replacement for Triton X-100 included the following: meeting global disposal requirements, effective virus inactivation, and no impact to product quality or process performance. Examples of detergents that were screened for inactivation of XMuLV are shown in Tables IV and V. The 1% Polysorbate 80 solution outperformed the 1% Polysorbate 20 solution with respect to XMuLV inactivation in the absence of the solvent, TnBP, with complete inactivation by Polysorbate 80 after 3 h (Table IV). Complete inactivation was achieved with both polysorbate solutions after only 1 h once the TnBP was added. However, substantial turbidity and increase in product aggregation was observed with both polysorbate/TnBP solutions (data not shown). Effective inactivation of XMuLV was also achieved after 1 h with the alkyl glucosides, sodium caprylate, and other vendor-proprietary detergents.
Inactivation of XMuLV in HCCF of MAb 1 with Detergents ± Solvent at 20 °C
Inactivation of XMuLV in HCCF of MAb 2 with Detergents at 20 °C
The impact of detergent exposure on product quality was assessed by incubation with 1% detergent in HCCF at ambient temperature overnight and subsequent purification on protein A. Comparable yields and chromatograms were observed on protein A chromatography. Size variants were measured with size exclusion chromatography and charge variants were measured with imaged capillary isoelectric focusing. As shown in Table VI, no product quality impact was noted for any of the detergents tested except for Lutensol XP90, where a slight increase in acidic variants was observed. No changes in low-molecular-weight species or process impurity levels were observed (data not shown).
Product Quality Impact of 1% (v/v) Detergent Incubation in HCCF for 18 h at Ambient Temperature
Virus Inactivation Robustness (Lower Concentration and Temperature)
Robustness of XMuLV inactivation was tested with known worse-case conditions of lower temperature and lower detergent concentration. As presented in Table VII, effective inactivation was achieved for each tested candidate at 12 °C and 0.3% detergent. To further understand the impact of detergent concentration, Triton CG110 was tested at various concentrations from 0.05% to 0.4%. Data summarized in Table VIII shows ineffective XMuLV inactivation with Triton CG110 at a concentration of 0.05%, wherein less than 1 LRV was obtained.
Inactivation of XMuLV in MAb 2 and rProtein 1 HCCF at Lower Temperature (12 °C) and Lower Concentration (0.3%, 0.5%) for 1 h
Inactivation of XMuLV in rProtein 2 HCCF at 12 °C and Various Concentrations of Triton CG110 for 1 hour
Virus Inactivation of Additional Enveloped Viruses
In addition to XMuLV, inactivation of pseudorabies virus (PRV) and bovine viral diarrhea virus (BVDV) was assessed with 0.3% detergents at 12 °C. As summarized in Table IX, the Ecosurf EH-9 and Lutensol XP90 detergents were not effective for inactivation of PRV, while the alkyl glucosides (APG 325N and Triton CG110) were effective for inactivation of all three tested viruses. As shown in Table X, Triton CG110 has comparable and effective inactivation of PRV and BVDV similar to Triton X-100.
Inactivation of Additional Enveloped Viruses in rProtein 1 HCCF with 0.3% Detergent at 12 °C for 1 h
Inactivation of Additional Enveloped Viruses in rProtein 2 HCCF with 0.3% Triton X-100 or Triton CG110 at 20 °C
Additional Considerations for Detergent Selection
Manufacturing process impacts such as detergent raw material stability, clearance during purification, process yields, foaming, and disposal were also considered. Several detergents were eliminated from consideration due to the formation of peroxides in the detergent raw material and aqueous stock solutions. Each of the top detergents cleared to ≤2 ng/mL in the protein A pools, as tested by nuclear magnetic resonance. No impact on purification process yield was observed for the tested detergents. Higher foaming levels were noted for the alkyl glucosides during addition to HCCF and processing; however, this may be mitigated with the use of an antifoaming agent if necessary. An assessment was performed on the top detergents to determine disposability of a 1% (v/v) detergent working concentration. There were some restrictions on disposal of each of the top candidates except for Triton CG110, which was acceptable for discharge at many manufacturing sites. Eco-friendly alkyl glucosides such as Triton CG110 provide robust inactivation of enveloped viruses while maintaining product quality and avoiding potential disposal issues encountered with octylphenol ethoxylates.
XMuLV Inactivation By Triton X-100: Impact of Process Conditions and Points of Use (Shivanthi Chinniah and Lisa Connell-Crowley, Amgen)
Triton X-100 is a detergent commonly used as a virus inactivation step in MAb purification processes. Triton X-100 is typically added to harvested cell culture fluid or the protein A elution pool and held for up to 60 min to inactivate enveloped viruses. In this study, the effect of Triton X-100 concentration, hold time, and harvest material composition on the inactivation of XMuLV was examined (Table XI). In cell culture medium in the presence of 0.1% Triton X-100, complete inactivation of XMuLV was achieved in 30 seconds. In HCCF, inactivation was rapid but 0.3% Triton X-100 was required to achieve complete inactivation. The increase in the Triton X-100 concentration required for complete inactivation is likely due to the presence of host cell impurities including cell membrane components, which may bind Triton X-100. Increasing the level of host cell impurities further by lysing the cells before harvest clarification resulted in incomplete inactivation at 0.3% Triton X-100 for two different molecules at either 30 seconds or 3 min (Table XII). These data indicate that Triton X-100 inactivation can occur very rapidly (<30 seconds) under appropriate conditions, but the amount of host cell impurities can affect the minimum concentration of Triton X-100 needed to achieve complete inactivation.
XMuLV Inactivation with 0.1 or 0.3% Triton X-100 at 15 °C
Effect of Different Harvest Materials on XMuLV Inactivation with 0.3% Triton X-100 at 15 °C
Summary
Low-pH Inactivation
The low-pH viral inactivation study presented at the 2013 symposium concluded that an acetate buffer system (compared to a glycine or citrate buffer system) was the most robust, enabling complete inactivation across a pH range (pH 3.3 to 3.9) regardless of temperature and protein concentration. Sodium acetate buffer was, therefore, deemed optimal for inactivation of XMuLV. The inactivation of XMuLV in the other two buffer systems is robust at pH 3.3; however, inactivation at higher pH values is more difficult to predict. Although other companies generally confirmed the trends as determined in the above presentation, their results were not altogether congruent with each other. Namely, a more robust viral inactivation at pH 3.6 with citrate buffer has been observed. Further investigations are, therefore, required to better understand the differences in buffer systems and to establish whether citrate buffer is problematic and acetate buffer is advantageous, as well as to measure the effect of varying buffer molarities on virus LRVs.
Detergent Inactivation
Because Triton X-100 is toxic to aquatic organisms and is subject to disposal restrictions at some sites, there is a need to find alternative detergents for viral inactivation during biopharmaceutical manufacturing. In the first presentation, the zwitterionic detergent LDAO was selected for further optimization from 30 potential candidates based on cost, low ecotoxicity, and its rapid inactivation of XMuLV and PRV in HCCF containing a recombinant Fc fusion protein. At an optimal concentration of 0.138% (w/v), LDAO was effective at rapidly inactivating XMuLV and Suid herpesvirus in HCCF for three MAbs and one Fc fusion protein with no detectable impact on quality attributes, including biological activity. Also, LDAO was shown to be effectively removed by the subsequent protein A capture step.
The second presentation also discussed the evaluation of potential alternatives to Triton X-100 for effective viral inactivation in HCCF. Of the 10 detergents screened, two of the alkyl glucosides (0.3% APG 325N and 0.3% Triton CG110) were able to inactivate all three enveloped viruses tested (XMuLV, PRV, and BVDV) in HCCF containing a recombinant protein after only 1 hour at 12 °C. Both detergents were effectively cleared during subsequent protein A purification, and they had no impact on the product quality with respect to yields for aggregates, fragments, and charge variants. Triton CG110, however, is considered to be more eco-friendly than APG 325N and does not face restrictions on disposability.
In the third presentation, a study of Triton X-100 inactivation kinetics revealed that the time to XMuLV inactivation in HCCF and lysed cell supernatant is quick, regardless of concentration (0.1% or 0.3%). However, the amount of host cell impurities can affect the minimum concentration of Triton X-100 needed to achieve complete inactivation.
During the discussion of effective detergent concentrations after these presentations, it was stated that generally the industry standard has been approximately 1% (for 1 h). Based on the studies presented, however, there seems to be potential for reducing levels of Triton X-100 to below this industry standard. Some companies reported having validated anywhere from 0.4% to 0.8% for XMuLV. In the end, a useful goal would be to determine a critical level of Triton X-100 (for example, 0.5%) and link to an ASTM standard.
Additional future work should include investigating the inactivation kinetics of Triton X-100 side-by-side with newly identified alternative candidate detergents such as LDAO, APG 325N, and others discussed above. Also, it would be important to characterize the robustness of inactivation of low-concentration Triton X-100 and alternative candidate detergents to a wider range of viruses with varying properties (e.g., more highly resistant viruses). The potential impact of rapid inactivation kinetics, as seen with Triton X-100 and alternative candidate detergents, on study duration would need to be assessed.
Although SB and LDAO displayed low potential for human harm, this has yet to be established to the same degree for Triton CG110 and other candidate detergents. Despite the fact that these detergents are used fairly upstream during purification, their consistent removal from the final drug substance and/or drug product should be demonstrated via in-process testing and/or final drug substance testing. It is advised that methods for measuring residual detergents be properly validated, including the establishment of limits of detection and limits of quantitation. In addition, as part of the body of evidence for characterizing the potential for human harm, the tolerable daily exposure values for any new candidate detergents should be determined.
- © PDA, Inc. 2015
References
Jump to section
- Article
- Introduction
- Virus Inactivation At Low Ph: Impact of Buffers, Ph Value, Protein Concentration, and Temperature On the Inactivation of Retroviruses (Franz Nothelfer, Hans-Georg Bischof, Helmut Winter, Joey Studts, Boehringer Ingelheim Pharma Gmbh & Co. Kg; Alice Marinaci and Horst Ruppach, Charles River Biopharmaceutical Services)
- Method of Inactivating Enveloped Viruses Using Eco-friendly, Cost Effective, Zwitterionic Detergents (Lynn Conley, Biogen Idec)
- Investigation of Alternative and Eco-friendly Detergents for Viral Inactivation (S. Fisher and L. Norling, Genentech)
- XMuLV Inactivation By Triton X-100: Impact of Process Conditions and Points of Use (Shivanthi Chinniah and Lisa Connell-Crowley, Amgen)
- Summary
- References
- Figures & Data
- References
- Info & Metrics
Related Articles
Cited By...
- Using Sensitivity Analysis to Simplify the Virus Safety Factor Calculation in the Manufacture of Biopharmaceuticals
- The Use of Bayesian Hierarchical Logistic Regression in the Development of a Modular Viral Inactivation Claim
- Session 1.1: Viral Clearance Using Traditional, Well-Understood Unit Operations: Low pH and Detergent Viral Inactivation