Downloaded from journal.pda.org on October 18, 2015
Viral Clearance Using Traditional, Well-Understood Unit Operations (Session I): Low-pH Inactivation Qi Chen
PDA J Pharm Sci and Tech 2014, 68 17-22 Access the most recent version at doi:10.5731/pdajpst.2014.00962
Downloaded from journal.pda.org on October 18, 2015
CONFERENCE PROCEEDING – ARTICLE
Viral Clearance Using Traditional, Well-Understood Unit Operations (Session I): Low-pH Inactivation QI CHEN, Ph.D. Genentech, a Member of the Roche Group; Process Virology, 1 DNA Way, Mailstop 10, South San Francisco, CA 94080 ©PDA, Inc. 2014 Background Extensive data from regulatory agency and industry were presented and discussed at the 2009 symposium (1). The collective data supported the effectiveness of the low-pH retrovirus inactivation, and process parameter ranges ensuring effective and reproducible viral clearance were defined. Four presentations at the 2011 symposium described studies to further understand critical process parameters and their interactions, as well as important factors in low-pH viral inactivation study design and execution to analyze outlier results. Development of a Low-pH Viral Inactivation (VI) Model and Analysis of an Outlier (Lisa Connell-Crowley, Shivanthi Chinniah, and Pete Hinckley; Amgen) A statistically designed fractional factorial experiment was used to examine the impact of several parameters on murine leukemia virus (MuLV) inactivation kinetics including pH, temperature, inactivation time, pretitration acetate concentration, and acid titrant (Table I). Runs were performed using monoclonal antibodies (mAbs) formulated at 20 g/L in the desired pre-titration acetate concentration, prior to low-pH titration with the acid titrant at the desired temperature. The design consisted of 18 runs over 6 days, with one midpoint run titrated with acetic acid performed each day to evaluate experimental variability. All other runs were performed at the upper and lower bounds of the experimental values, with the exception of acid titrant. Figure 1 shows the inactivation curves for each run at the different pH values. At pH 3.6, inactivation was rapid with no virus detectible after 10 min in all runs, indicating that variations in the other parameters had little impact. The pH 3.7 runs, which were all center-
doi: 10.5731/pdajpst.2014.00962
Vol. 68, No. 1, January-February 2014
point runs, show somewhat slower inactivation kinetics but were relatively consistent, with five out of six runs showing no detectible virus at the 60 min time point. At pH 3.8, the inactivation curves show much slower inactivation and no runs achieved complete inactivation at the 90 min time point. The variability in kinetics for the pH 3.8 runs also demonstrates that other parameters besides pH were affecting the inactivation kinetics. The results were used to develop models for low-pH inactivation at the different time points (Figure 2). The 0.5, 3, and 6 min time points had the best model fit with R2 values above 0.9. The models for the 10 and 60 min time points did not fit as well (R2 of 0.55 and 0.24, respectively) due to the fact that a number of the runs had reached complete inactivation and thus had similar values. As expected, pH was the most significant contributor to inactivation kinetics. Temperature and pre-titration acetate concentration had a lesser but significant impact, with slower kinetics observed at lower temperature and lower acetate concentrations. The impact of acetate concentration is similar to that previously demonstrated with glycine, where lower glycine concentrations negatively affected inactivation kinetics (1), suggesting ionic strength rather than buffer type is important. The titrant acetic acid also appeared to slightly improve inactivation kinetics as compared to phosphoric or formic acid. This effect is attributed to the fact that acetic acid is a weaker than formic or phosphoric acid and thus more ions are added during an acetic acid titration, resulting in an increase in the overall ionic strength of the pool. Based on these data and building from earlier work (2), pH 3.6 is predicted to provide the most robust and rapid inactivation over the range evaluated (pH 3.6 to 3.8), as the effects of temperature, pre-titration acetate concentration, and acid titrant are no longer practically significant. At pH 3.7 and 3.8, the models can predict whether certain operating conditions will negatively affect inactivation kinetics. If lowering the pH to 3.6 is 17
Downloaded from journal.pda.org on October 18, 2015
Table I Factors and Set Points Examined by an Amgen DOE Study of Low-pH Inactivation of X-MuLV. The DoE Used a Fractional Factorial, D-Optimal Design Factor
Values
pH
3.6, 3.7, 3.8
Temperature (°C)
15, 18, 21
Pre-titration acetate (mM)
25, 62.5, 100
Acid titrant
formic, acetic, phosphoric
not feasible, improvements might be achieved by higher temperature or an increase in ionic strength from either the starting buffer concentration or by choosing a weaker acid such as acetic acid.
Figure 2 Amgen model of the DoE outlined in Table I evaluating the impact of each parameter on X-MuLV inactivation. Analysis of the log reduction value at the 3 and 6 min time points as predicted by the model is shown. Similar results were observed for the 0.5 min time point. Data courtesy of Lisa Connell-Crowley, Shivanthi Chinniah, and Pete Hinckley.
Defining Robust Operating Ranges for Low-pH Inactivation in a mAb Platform Process for Reproducible Viral Clearance (D. Vacante, H. Shen, and T. Cheung; Development, Janssen Pharmaceuticals)
Figure 1 Amgen study of the inactivation kinetics of X-MuLV at pH 3.6, 3.7 (centerpoint), and 3.8. Virus titer was measured by TCID50 at 0.5, 3, 6, 10, and 60 min. Open symbols indicate no virus was detected at that time point. Data courtesy of Lisa Connell-Crowley, Shivanthi Chinniah, and Pete Hinckley. 18
A low-pH inactivation unit operation is typically used in biopharmaceutical manufacturing processes as a dedicated step to inactivate retroviruses, other enveloped viruses, and other viruses sensitive to low pH. Janssen showed inactivation to be robust below pH 3.8; however, when operating at or above pH 3.8, the inactivation may be influenced by temperature and concentration of antibody. Janssen also showed that the spiking virus can influence the small-scale inactivation if not properly controlled. Because the spiking virus preparation is in a neutral pH buffer and there may be weak buffering capacity of the low-pH treatment process pool, small-scale studies may be affected when virus is added to the process pool. The data in Table II show that the pH of the antibody process pool increases when adding the virus formulation buffer or media at typical spiking ratios at various antibody concentrations. The effect is slightly more pronounced at lower antibody concentrations, most likely due to the weaker buffering capacity of these process pools. In studies using MuLV, the log reduction value (LRV) was 1.5–2.0 after 60 min of incubation time using similar conditions described in Table II (data not shown). This is most likely an experimental artifact due to an increase in pH, as when the pH was conPDA Journal of Pharmaceutical Science and Technology
Downloaded from journal.pda.org on October 18, 2015
Table II Subtle Shifts in the pH of a Janssen Monoclonal Antibody Process Pool at Various Protein Concentrations after Addition of Spiking Virus Media. Data Courtesy of Pedro Alfonso and Tiffany Cheung pH Measurement at Various mAb Concentrations Test Condition
4 mg/mL
9 mg/mL
11 mg/mL
20 mg/mL
Start pH of DPC eluate pH @ Room Temp after media addition pH @ 19 °C after media addition pH @ 18 °C after media addition
3.81 3.96 N/A 4.09
3.86 3.98 N/A 4.03
3.69 3.80 3.83 3.87
3.80 3.86 N/A 3.98
trolled with regard to the spiking virus at pH 3.8, the LRV was in the range of 5.5– 6.0 after 60 min of incubation time. In a design of experiment (DoE) study with one particular antibody, 12 inactivation runs were performed to evaluate the proven acceptable ranges for low-pH inactivation for temperature, pH, and antibody concentration. The spike ratio and inactivation times were fixed in the study. The study design is provided in Table III. The LRV for MuLV from the DoE study were plotted versus temperature, pH, and concentration, and interaction profiles were generated (provided in Figure 3). At pH 3.5, the LRV did not vary significantly with a change in temperature or concentration (Figure 3, middle row). However, at pH 3.9, the slope of the LRV line increases, indicating an effect of temperature and concentration on LRV or inactivation (Figure 3, middle row). When pH was varied (Figure 3, middle column), curvature was observed for the temperature and antibody protein concentration profiles, indicating what may be an exponential decay in LRV, possibly indicating the edge of failure for inactivation.
In addition, Janssen provided a statistical analysis of its viral clearance database with approximately 100 data points for more than five antibodies, comparing LRV using contour plots for temperature vs pH and concentration vs pH. In this analysis, robust operating regions could be defined; when pH is 3.5–3.9 with an antibody concentration of 10 – 40 g/L and at a temperature of 18 –25 °C, worst-case conditions for all parameters would provide at least 4.5 log10 clearance of MuLV (Figure 4). In conclusion, when performing small-scale, low-pH inactivation studies it is important to control and monitor the pH (post-viral spike) so that it is at the expected experimental condition. In addition, the DoE approach was useful to study interactions of the various operating parameters. The DoE study indicated that pH 3.9 represents the borderline of robustness for the low-pH inactivation conditions. At this pH, vary-
Table III Janssen DOE Study Design for Low-pH Inactivation of MuLV. Data Courtesy of Al Magill. 1
2
DOE Factor
Temp (°C)
pH
3 Protein Concentration (g/L)
Expected PAR Level 1 Level 2 Level 3
17–25 15 25 N/A
3.5–3.8 3.5 3.7 3.9
10–40 5 40 N/A
Vol. 68, No. 1, January-February 2014
Figure 3 Interaction profiles for the Janssen MuLV inactivation DOE study (study design in Table III). Data courtesy of Al Magill with analysis by Hong Shen. 19
Downloaded from journal.pda.org on October 18, 2015
Figure 4 Response surface analysis of Janssen’s cumulative low-pH inactivation experience with five antibodies (more than 100 data points). (a) Left panel is temperature (°C) versus pH. (b) Right panel is monoclonal antibody concentration (g/L) versus pH. LRV for each condition is presented as a contour with various colors for a LRV range of 2.5– 6.5 (panel a and b), with blue representing low end of the LRV range and red representing the high end. The green box within each plot is the area with clearance greater than approximately 4.5. There is an exception in panel b where a small area in the lower right section of the green box has a LRV below 4.5 due to being run at 15 °C, a temperature outside of the proven acceptable range. Data analysis courtesy of Hong Shen. ing antibody concentrations, temperature, spike ratio, and time can have an impact on the LRV. Lastly, review and analysis of our viral clearance database indicates that worst-case conditions provide at least 4.5 log10 of MuLV clearance. Virus Inactivation at Low pH (S. Kobayashi, Chugai Pharmaceuticals) Low-pH treatment is widely used in the downstream process as an effective virus inactivation unit operation. However, antibody aggregation and fragmentation are sometimes observed during low-pH inactivation. Determining the limit and evaluating its impact of critical process parameters would contribute to decreasing the aggregate and/or fragment formation and increasing the product quality and yield.
In addition, data with wide range of isoelectric point (pI) value mAbs (5.7–9.4) were collected to produce a better understanding of a “bracketed generic clearance” on low-pH inactivation, proposed by Brorson et al. (2). As shown in Figure 6, LRVs were approximately 6 or more than 6 log10 at pH 3.9 for 60 min without significant differences. However, only in the case of mAb4 (pI 5.7), the LRV for 30 min was relatively lower than the others. The LRVs didn’t seem to be associated with each pI value, so the lower LRV on mAb4 was probably not caused by pI value.
In the study, Chugai evaluated whether operating temperature would be a critical process parameter, and how much it would have an impact on LRV. MuLV-spiked mAbs eluted with 50 mmol/L acetic acid or 2.5 mmol/L HCl from protein A resin, approximately 15 mg/mL, were incubated at pH 3.9 at each temperature, 15, 20 and 25 °C, for 5, 30, and 60 min. The result revealed, as shown in Figure 5, that for every 5 °C rise in incubation temperature the LRV increase approximately 1 log10. The comparable data was obtained with 2.5 mmol/L HCl (data not shown). 20
Figure 5 Chugai study of clearance of X-MuLV by low-pH treatment at tested temperature, 15, 20 and 25 °C. Spike virus concentration was approximately 7 log10 TCID50/mL, and spiked ratio was 10%. Accuracy of temperature was ⴞ0.3 °C, and pH was ⴞ0.05. Data was collected in a duplicate manner. PDA Journal of Pharmaceutical Science and Technology
Downloaded from journal.pda.org on October 18, 2015
Figure 6 Chugai study of clearance of X-MuLV by low-pH treatment using wide range of pI value-mAbs (pI 5.7⬃9.4). Spike virus concentration was approximately 7 log10 TCID50/mL, and spiked ratio was 10%. Accuracy of temperature was ⴞ0.3 °C, and pH was ⴞ0.05. Data was collected in a duplicate manner.
To put it simply, pI value may not be critical for virus inactivation, but our data indicates treating not less than 60 min is recommended for robust viral clearance, at 25 °C and pH 3.9. Low pH Viral Inactivation: We’re Not Done Yet (Daniel M. Strauss and Dayue Chen, Eli Lilly and Company) Low-pH treatment, as one of the dedicated unit operations for virus clearance at Eli Lilly and Company, has been providing robust retroviral clearance for
many different products provided that the operating ranges of several critical process parameters are appropriately controlled. Specifically, conditions of pH ⱕ3.65 ⫾ 0.05, ambient temperature (⬃20 °C), and 1–2 h treatment have consistently provided effective retrovirus clearance and often eliminate the virus to below the limit of detection (complete clearance). However, it is not uncommon that low-pH viral inactivation has to be evaluated at elevated pH (e.g., 3.85 ⫾ 0.05) to minimize the adverse impact of low pH on product quality and stability. Recent data obtained in our laboratory suggests that for some products, the effectiveness of viral inactivation at elevated pH can be highly sensitive to treatment temperature, whereas other products display no such sensitivity. For instance, at pH ⬃3.85, mAb A showed comparable inactivation kinetics at 15 °C and 25 °C and LRV was largely not affected by temperature differences (Figure 7). In contrast, inactivation kinetics for mAb B exhibited significant temperature sensitivity as the 60 min treatment achieved LRVs of 2.9 log10 and 6.0 log10 at 15 °C and 25 °C, respectively (Figure 7). Similar temperature-dependent inactivation was observed for another molecule (mAb C) as shown in Table IV, indicating that temperature could potentially influence the inactivation kinetics as well as the overall LRV at elevated pH. Based on these results, we conclude that inactivation robustness by low-pH treatment at elevated pH (e.g., 3.8) must be carefully evaluated on a process/product specific basis
Figure 7 Comparison of retrovirus inactivation kinetics by low-pH treatment at 3.85 ⴞ 0.05 for different Lilly products at two different temperatures. LRF, logarithmic reduction factor. Vol. 68, No. 1, January-February 2014
21
Downloaded from journal.pda.org on October 18, 2015
Table IV Effects of Temperature on Low-pH Retrovirus Inactivation in the Presence of Four Lilly Process Intermediates. LRF, Log10 Reduction Factor LRF Achieved by 60 min Treatment at Indicated pH and Temperature Product
Isotype
Buffer
pH: ⬃3.65 T ⴝ Ambient
mAb A mAb B mAb C Protein X
IgG1 IgG1 IgG4 Fc-Fusion
Succinate Citrate Citrate Citrate
⬎6.6 5.7 ⬎5.3 ⬎5.4
to ensure product safety and effective control at manufacturing. Summary Data presented confirmed that low-pH viral inactivation is a well-understood step with reproducible data from the industry. The most critical parameter is pH, where pH ⬍3.6 is robust in achieving retrovirus inactivation across mAbs and process parameter ranges. At higher pH, parameters such as temperature, product concentration, affinity column elution buffer, acid titrant, as well as mAb may affect viral inactivation. The participants concluded that high pH, short time, low temperature, and low ionic strength (at least for acetate and glycine buffers) are considered worst-case conditions for virus inactivation. The impact of product concentration is somewhat controversial. Higher protein concentration was assumed to be a worst-case scenario. However, Janssen’s DoE experiment indicated that at pH 3.9, lower protein concentration is correlated with lower LRVs. More experiments are needed to confirm this observation (e.g., impact of protein concentration and time on inactivation kinetics). No difference in viral inactivation has been observed relating to the strain of retrovirus or the properties of the virus spike. Two firms presented analysis of outliers. It was found that pH probe health is critical to obtain accurate pH measurement (Amgen, data not shown). Janssen re-
22
pH: ⬃3.85 T ⴝ 15 °C
pH: ⬃3.85 T ⴝ 25 °C
⬎6.4 2.9 3.4 ⬎5.3
⬎6.5 6.0 ⬎4.3 ⬎5.0
ported that pH could increase after virus spike and leads to lower-than-expected viral inactivation. Because a low protein concentration pool may have lower buffering capacity, spiking with virus at neutral pH could result in a pool pH higher than the target. Thus the lower LRV could be mistakenly attributed to low protein concentration. In order to obtain valid results, it is important to understand the impact of the virus spike on the resultant pH of the solution and to control these factors carefully during a study. Based on the data presented here and a host of historical observations, an effort was formed at the meeting to write a protocol for performing low-pH validation studies. This effort was spearheaded by Dominick Vacante of Janssen, and the work product is presented in this issue of the PDA Journal. References 1. Miesegaes, G.; Bailey, M.; Willkommen, H.; Chen, Q.; Roush, D.; Blu¨mel, J.; Brorson, K. Proceedings of the 2009 Viral Clearance Symposium. Dev. Biol. (Basel) 2010, 133, 3–101. 2. Brorson, K.; Krejci, S.; Lee, K.; Hamilton, E.; Stein, K.; Xu, Y. Bracketed generic inactivation of rodent retroviruses by low pH treatment for monoclonal antibodies and recombinant proteins. Biotechnol. Bioeng. 2003, 82 (3), 321–329.
PDA Journal of Pharmaceutical Science and Technology
Downloaded from journal.pda.org on October 18, 2015
An Authorized User of the electronic PDA Journal of Pharmaceutical Science and Technology (the PDA Journal) is a PDA Member in good standing. Authorized Users are permitted to do the following: ·Search and view the content of the PDA Journal ·Download a single article for the individual use of an Authorized User ·Assemble and distribute links that point to the PDA Journal ·Print individual articles from the PDA Journal for the individual use of an Authorized User ·Make a reasonable number of photocopies of a printed article for the individual use of an Authorized User or for the use by or distribution to other Authorized Users Authorized Users are not permitted to do the following: ·Except as mentioned above, allow anyone other than an Authorized User to use or access the PDA Journal · Display or otherwise make any information from the PDA Journal available to anyone other than an Authorized User ·Post articles from the PDA Journal on Web sites, either available on the Internet or an Intranet, or in any form of online publications ·Transmit electronically, via e-mail or any other file transfer protocols, any portion of the PDA Journal ·Create a searchable archive of any portion of the PDA Journal ·Use robots or intelligent agents to access, search and/or systematically download any portion of the PDA Journal ·Sell, re-sell, rent, lease, license, sublicense, assign or otherwise transfer the use of the PDA Journal or its content ·Use or copy the PDA Journal for document delivery, fee-for-service use, or bulk reproduction or distribution of materials in any form, or any substantially similar commercial purpose ·Alter, modify, repackage or adapt any portion of the PDA Journal ·Make any edits or derivative works with respect to any portion of the PDA Journal including any text or graphics ·Delete or remove in any form or format, including on a printed article or photocopy, any copyright information or notice contained in the PDA Journal
Viral Clearance Using Traditional, Well-Understood Unit Operations (Session I): Low-pH Inactivation Qi Chen
PDA J Pharm Sci and Tech 2014, 68 17-22 Access the most recent version at doi:10.5731/pdajpst.2014.00962
Downloaded from journal.pda.org on October 18, 2015
CONFERENCE PROCEEDING – ARTICLE
Viral Clearance Using Traditional, Well-Understood Unit Operations (Session I): Low-pH Inactivation QI CHEN, Ph.D. Genentech, a Member of the Roche Group; Process Virology, 1 DNA Way, Mailstop 10, South San Francisco, CA 94080 ©PDA, Inc. 2014 Background Extensive data from regulatory agency and industry were presented and discussed at the 2009 symposium (1). The collective data supported the effectiveness of the low-pH retrovirus inactivation, and process parameter ranges ensuring effective and reproducible viral clearance were defined. Four presentations at the 2011 symposium described studies to further understand critical process parameters and their interactions, as well as important factors in low-pH viral inactivation study design and execution to analyze outlier results. Development of a Low-pH Viral Inactivation (VI) Model and Analysis of an Outlier (Lisa Connell-Crowley, Shivanthi Chinniah, and Pete Hinckley; Amgen) A statistically designed fractional factorial experiment was used to examine the impact of several parameters on murine leukemia virus (MuLV) inactivation kinetics including pH, temperature, inactivation time, pretitration acetate concentration, and acid titrant (Table I). Runs were performed using monoclonal antibodies (mAbs) formulated at 20 g/L in the desired pre-titration acetate concentration, prior to low-pH titration with the acid titrant at the desired temperature. The design consisted of 18 runs over 6 days, with one midpoint run titrated with acetic acid performed each day to evaluate experimental variability. All other runs were performed at the upper and lower bounds of the experimental values, with the exception of acid titrant. Figure 1 shows the inactivation curves for each run at the different pH values. At pH 3.6, inactivation was rapid with no virus detectible after 10 min in all runs, indicating that variations in the other parameters had little impact. The pH 3.7 runs, which were all center-
doi: 10.5731/pdajpst.2014.00962
Vol. 68, No. 1, January-February 2014
point runs, show somewhat slower inactivation kinetics but were relatively consistent, with five out of six runs showing no detectible virus at the 60 min time point. At pH 3.8, the inactivation curves show much slower inactivation and no runs achieved complete inactivation at the 90 min time point. The variability in kinetics for the pH 3.8 runs also demonstrates that other parameters besides pH were affecting the inactivation kinetics. The results were used to develop models for low-pH inactivation at the different time points (Figure 2). The 0.5, 3, and 6 min time points had the best model fit with R2 values above 0.9. The models for the 10 and 60 min time points did not fit as well (R2 of 0.55 and 0.24, respectively) due to the fact that a number of the runs had reached complete inactivation and thus had similar values. As expected, pH was the most significant contributor to inactivation kinetics. Temperature and pre-titration acetate concentration had a lesser but significant impact, with slower kinetics observed at lower temperature and lower acetate concentrations. The impact of acetate concentration is similar to that previously demonstrated with glycine, where lower glycine concentrations negatively affected inactivation kinetics (1), suggesting ionic strength rather than buffer type is important. The titrant acetic acid also appeared to slightly improve inactivation kinetics as compared to phosphoric or formic acid. This effect is attributed to the fact that acetic acid is a weaker than formic or phosphoric acid and thus more ions are added during an acetic acid titration, resulting in an increase in the overall ionic strength of the pool. Based on these data and building from earlier work (2), pH 3.6 is predicted to provide the most robust and rapid inactivation over the range evaluated (pH 3.6 to 3.8), as the effects of temperature, pre-titration acetate concentration, and acid titrant are no longer practically significant. At pH 3.7 and 3.8, the models can predict whether certain operating conditions will negatively affect inactivation kinetics. If lowering the pH to 3.6 is 17
Downloaded from journal.pda.org on October 18, 2015
Table I Factors and Set Points Examined by an Amgen DOE Study of Low-pH Inactivation of X-MuLV. The DoE Used a Fractional Factorial, D-Optimal Design Factor
Values
pH
3.6, 3.7, 3.8
Temperature (°C)
15, 18, 21
Pre-titration acetate (mM)
25, 62.5, 100
Acid titrant
formic, acetic, phosphoric
not feasible, improvements might be achieved by higher temperature or an increase in ionic strength from either the starting buffer concentration or by choosing a weaker acid such as acetic acid.
Figure 2 Amgen model of the DoE outlined in Table I evaluating the impact of each parameter on X-MuLV inactivation. Analysis of the log reduction value at the 3 and 6 min time points as predicted by the model is shown. Similar results were observed for the 0.5 min time point. Data courtesy of Lisa Connell-Crowley, Shivanthi Chinniah, and Pete Hinckley.
Defining Robust Operating Ranges for Low-pH Inactivation in a mAb Platform Process for Reproducible Viral Clearance (D. Vacante, H. Shen, and T. Cheung; Development, Janssen Pharmaceuticals)
Figure 1 Amgen study of the inactivation kinetics of X-MuLV at pH 3.6, 3.7 (centerpoint), and 3.8. Virus titer was measured by TCID50 at 0.5, 3, 6, 10, and 60 min. Open symbols indicate no virus was detected at that time point. Data courtesy of Lisa Connell-Crowley, Shivanthi Chinniah, and Pete Hinckley. 18
A low-pH inactivation unit operation is typically used in biopharmaceutical manufacturing processes as a dedicated step to inactivate retroviruses, other enveloped viruses, and other viruses sensitive to low pH. Janssen showed inactivation to be robust below pH 3.8; however, when operating at or above pH 3.8, the inactivation may be influenced by temperature and concentration of antibody. Janssen also showed that the spiking virus can influence the small-scale inactivation if not properly controlled. Because the spiking virus preparation is in a neutral pH buffer and there may be weak buffering capacity of the low-pH treatment process pool, small-scale studies may be affected when virus is added to the process pool. The data in Table II show that the pH of the antibody process pool increases when adding the virus formulation buffer or media at typical spiking ratios at various antibody concentrations. The effect is slightly more pronounced at lower antibody concentrations, most likely due to the weaker buffering capacity of these process pools. In studies using MuLV, the log reduction value (LRV) was 1.5–2.0 after 60 min of incubation time using similar conditions described in Table II (data not shown). This is most likely an experimental artifact due to an increase in pH, as when the pH was conPDA Journal of Pharmaceutical Science and Technology
Downloaded from journal.pda.org on October 18, 2015
Table II Subtle Shifts in the pH of a Janssen Monoclonal Antibody Process Pool at Various Protein Concentrations after Addition of Spiking Virus Media. Data Courtesy of Pedro Alfonso and Tiffany Cheung pH Measurement at Various mAb Concentrations Test Condition
4 mg/mL
9 mg/mL
11 mg/mL
20 mg/mL
Start pH of DPC eluate pH @ Room Temp after media addition pH @ 19 °C after media addition pH @ 18 °C after media addition
3.81 3.96 N/A 4.09
3.86 3.98 N/A 4.03
3.69 3.80 3.83 3.87
3.80 3.86 N/A 3.98
trolled with regard to the spiking virus at pH 3.8, the LRV was in the range of 5.5– 6.0 after 60 min of incubation time. In a design of experiment (DoE) study with one particular antibody, 12 inactivation runs were performed to evaluate the proven acceptable ranges for low-pH inactivation for temperature, pH, and antibody concentration. The spike ratio and inactivation times were fixed in the study. The study design is provided in Table III. The LRV for MuLV from the DoE study were plotted versus temperature, pH, and concentration, and interaction profiles were generated (provided in Figure 3). At pH 3.5, the LRV did not vary significantly with a change in temperature or concentration (Figure 3, middle row). However, at pH 3.9, the slope of the LRV line increases, indicating an effect of temperature and concentration on LRV or inactivation (Figure 3, middle row). When pH was varied (Figure 3, middle column), curvature was observed for the temperature and antibody protein concentration profiles, indicating what may be an exponential decay in LRV, possibly indicating the edge of failure for inactivation.
In addition, Janssen provided a statistical analysis of its viral clearance database with approximately 100 data points for more than five antibodies, comparing LRV using contour plots for temperature vs pH and concentration vs pH. In this analysis, robust operating regions could be defined; when pH is 3.5–3.9 with an antibody concentration of 10 – 40 g/L and at a temperature of 18 –25 °C, worst-case conditions for all parameters would provide at least 4.5 log10 clearance of MuLV (Figure 4). In conclusion, when performing small-scale, low-pH inactivation studies it is important to control and monitor the pH (post-viral spike) so that it is at the expected experimental condition. In addition, the DoE approach was useful to study interactions of the various operating parameters. The DoE study indicated that pH 3.9 represents the borderline of robustness for the low-pH inactivation conditions. At this pH, vary-
Table III Janssen DOE Study Design for Low-pH Inactivation of MuLV. Data Courtesy of Al Magill. 1
2
DOE Factor
Temp (°C)
pH
3 Protein Concentration (g/L)
Expected PAR Level 1 Level 2 Level 3
17–25 15 25 N/A
3.5–3.8 3.5 3.7 3.9
10–40 5 40 N/A
Vol. 68, No. 1, January-February 2014
Figure 3 Interaction profiles for the Janssen MuLV inactivation DOE study (study design in Table III). Data courtesy of Al Magill with analysis by Hong Shen. 19
Downloaded from journal.pda.org on October 18, 2015
Figure 4 Response surface analysis of Janssen’s cumulative low-pH inactivation experience with five antibodies (more than 100 data points). (a) Left panel is temperature (°C) versus pH. (b) Right panel is monoclonal antibody concentration (g/L) versus pH. LRV for each condition is presented as a contour with various colors for a LRV range of 2.5– 6.5 (panel a and b), with blue representing low end of the LRV range and red representing the high end. The green box within each plot is the area with clearance greater than approximately 4.5. There is an exception in panel b where a small area in the lower right section of the green box has a LRV below 4.5 due to being run at 15 °C, a temperature outside of the proven acceptable range. Data analysis courtesy of Hong Shen. ing antibody concentrations, temperature, spike ratio, and time can have an impact on the LRV. Lastly, review and analysis of our viral clearance database indicates that worst-case conditions provide at least 4.5 log10 of MuLV clearance. Virus Inactivation at Low pH (S. Kobayashi, Chugai Pharmaceuticals) Low-pH treatment is widely used in the downstream process as an effective virus inactivation unit operation. However, antibody aggregation and fragmentation are sometimes observed during low-pH inactivation. Determining the limit and evaluating its impact of critical process parameters would contribute to decreasing the aggregate and/or fragment formation and increasing the product quality and yield.
In addition, data with wide range of isoelectric point (pI) value mAbs (5.7–9.4) were collected to produce a better understanding of a “bracketed generic clearance” on low-pH inactivation, proposed by Brorson et al. (2). As shown in Figure 6, LRVs were approximately 6 or more than 6 log10 at pH 3.9 for 60 min without significant differences. However, only in the case of mAb4 (pI 5.7), the LRV for 30 min was relatively lower than the others. The LRVs didn’t seem to be associated with each pI value, so the lower LRV on mAb4 was probably not caused by pI value.
In the study, Chugai evaluated whether operating temperature would be a critical process parameter, and how much it would have an impact on LRV. MuLV-spiked mAbs eluted with 50 mmol/L acetic acid or 2.5 mmol/L HCl from protein A resin, approximately 15 mg/mL, were incubated at pH 3.9 at each temperature, 15, 20 and 25 °C, for 5, 30, and 60 min. The result revealed, as shown in Figure 5, that for every 5 °C rise in incubation temperature the LRV increase approximately 1 log10. The comparable data was obtained with 2.5 mmol/L HCl (data not shown). 20
Figure 5 Chugai study of clearance of X-MuLV by low-pH treatment at tested temperature, 15, 20 and 25 °C. Spike virus concentration was approximately 7 log10 TCID50/mL, and spiked ratio was 10%. Accuracy of temperature was ⴞ0.3 °C, and pH was ⴞ0.05. Data was collected in a duplicate manner. PDA Journal of Pharmaceutical Science and Technology
Downloaded from journal.pda.org on October 18, 2015
Figure 6 Chugai study of clearance of X-MuLV by low-pH treatment using wide range of pI value-mAbs (pI 5.7⬃9.4). Spike virus concentration was approximately 7 log10 TCID50/mL, and spiked ratio was 10%. Accuracy of temperature was ⴞ0.3 °C, and pH was ⴞ0.05. Data was collected in a duplicate manner.
To put it simply, pI value may not be critical for virus inactivation, but our data indicates treating not less than 60 min is recommended for robust viral clearance, at 25 °C and pH 3.9. Low pH Viral Inactivation: We’re Not Done Yet (Daniel M. Strauss and Dayue Chen, Eli Lilly and Company) Low-pH treatment, as one of the dedicated unit operations for virus clearance at Eli Lilly and Company, has been providing robust retroviral clearance for
many different products provided that the operating ranges of several critical process parameters are appropriately controlled. Specifically, conditions of pH ⱕ3.65 ⫾ 0.05, ambient temperature (⬃20 °C), and 1–2 h treatment have consistently provided effective retrovirus clearance and often eliminate the virus to below the limit of detection (complete clearance). However, it is not uncommon that low-pH viral inactivation has to be evaluated at elevated pH (e.g., 3.85 ⫾ 0.05) to minimize the adverse impact of low pH on product quality and stability. Recent data obtained in our laboratory suggests that for some products, the effectiveness of viral inactivation at elevated pH can be highly sensitive to treatment temperature, whereas other products display no such sensitivity. For instance, at pH ⬃3.85, mAb A showed comparable inactivation kinetics at 15 °C and 25 °C and LRV was largely not affected by temperature differences (Figure 7). In contrast, inactivation kinetics for mAb B exhibited significant temperature sensitivity as the 60 min treatment achieved LRVs of 2.9 log10 and 6.0 log10 at 15 °C and 25 °C, respectively (Figure 7). Similar temperature-dependent inactivation was observed for another molecule (mAb C) as shown in Table IV, indicating that temperature could potentially influence the inactivation kinetics as well as the overall LRV at elevated pH. Based on these results, we conclude that inactivation robustness by low-pH treatment at elevated pH (e.g., 3.8) must be carefully evaluated on a process/product specific basis
Figure 7 Comparison of retrovirus inactivation kinetics by low-pH treatment at 3.85 ⴞ 0.05 for different Lilly products at two different temperatures. LRF, logarithmic reduction factor. Vol. 68, No. 1, January-February 2014
21
Downloaded from journal.pda.org on October 18, 2015
Table IV Effects of Temperature on Low-pH Retrovirus Inactivation in the Presence of Four Lilly Process Intermediates. LRF, Log10 Reduction Factor LRF Achieved by 60 min Treatment at Indicated pH and Temperature Product
Isotype
Buffer
pH: ⬃3.65 T ⴝ Ambient
mAb A mAb B mAb C Protein X
IgG1 IgG1 IgG4 Fc-Fusion
Succinate Citrate Citrate Citrate
⬎6.6 5.7 ⬎5.3 ⬎5.4
to ensure product safety and effective control at manufacturing. Summary Data presented confirmed that low-pH viral inactivation is a well-understood step with reproducible data from the industry. The most critical parameter is pH, where pH ⬍3.6 is robust in achieving retrovirus inactivation across mAbs and process parameter ranges. At higher pH, parameters such as temperature, product concentration, affinity column elution buffer, acid titrant, as well as mAb may affect viral inactivation. The participants concluded that high pH, short time, low temperature, and low ionic strength (at least for acetate and glycine buffers) are considered worst-case conditions for virus inactivation. The impact of product concentration is somewhat controversial. Higher protein concentration was assumed to be a worst-case scenario. However, Janssen’s DoE experiment indicated that at pH 3.9, lower protein concentration is correlated with lower LRVs. More experiments are needed to confirm this observation (e.g., impact of protein concentration and time on inactivation kinetics). No difference in viral inactivation has been observed relating to the strain of retrovirus or the properties of the virus spike. Two firms presented analysis of outliers. It was found that pH probe health is critical to obtain accurate pH measurement (Amgen, data not shown). Janssen re-
22
pH: ⬃3.85 T ⴝ 15 °C
pH: ⬃3.85 T ⴝ 25 °C
⬎6.4 2.9 3.4 ⬎5.3
⬎6.5 6.0 ⬎4.3 ⬎5.0
ported that pH could increase after virus spike and leads to lower-than-expected viral inactivation. Because a low protein concentration pool may have lower buffering capacity, spiking with virus at neutral pH could result in a pool pH higher than the target. Thus the lower LRV could be mistakenly attributed to low protein concentration. In order to obtain valid results, it is important to understand the impact of the virus spike on the resultant pH of the solution and to control these factors carefully during a study. Based on the data presented here and a host of historical observations, an effort was formed at the meeting to write a protocol for performing low-pH validation studies. This effort was spearheaded by Dominick Vacante of Janssen, and the work product is presented in this issue of the PDA Journal. References 1. Miesegaes, G.; Bailey, M.; Willkommen, H.; Chen, Q.; Roush, D.; Blu¨mel, J.; Brorson, K. Proceedings of the 2009 Viral Clearance Symposium. Dev. Biol. (Basel) 2010, 133, 3–101. 2. Brorson, K.; Krejci, S.; Lee, K.; Hamilton, E.; Stein, K.; Xu, Y. Bracketed generic inactivation of rodent retroviruses by low pH treatment for monoclonal antibodies and recombinant proteins. Biotechnol. Bioeng. 2003, 82 (3), 321–329.
PDA Journal of Pharmaceutical Science and Technology
Downloaded from journal.pda.org on October 18, 2015
An Authorized User of the electronic PDA Journal of Pharmaceutical Science and Technology (the PDA Journal) is a PDA Member in good standing. Authorized Users are permitted to do the following: ·Search and view the content of the PDA Journal ·Download a single article for the individual use of an Authorized User ·Assemble and distribute links that point to the PDA Journal ·Print individual articles from the PDA Journal for the individual use of an Authorized User ·Make a reasonable number of photocopies of a printed article for the individual use of an Authorized User or for the use by or distribution to other Authorized Users Authorized Users are not permitted to do the following: ·Except as mentioned above, allow anyone other than an Authorized User to use or access the PDA Journal · Display or otherwise make any information from the PDA Journal available to anyone other than an Authorized User ·Post articles from the PDA Journal on Web sites, either available on the Internet or an Intranet, or in any form of online publications ·Transmit electronically, via e-mail or any other file transfer protocols, any portion of the PDA Journal ·Create a searchable archive of any portion of the PDA Journal ·Use robots or intelligent agents to access, search and/or systematically download any portion of the PDA Journal ·Sell, re-sell, rent, lease, license, sublicense, assign or otherwise transfer the use of the PDA Journal or its content ·Use or copy the PDA Journal for document delivery, fee-for-service use, or bulk reproduction or distribution of materials in any form, or any substantially similar commercial purpose ·Alter, modify, repackage or adapt any portion of the PDA Journal ·Make any edits or derivative works with respect to any portion of the PDA Journal including any text or graphics ·Delete or remove in any form or format, including on a printed article or photocopy, any copyright information or notice contained in the PDA Journal
