Assessing Viral Inactivation for Continuous Processing

The complexity of many biopharmaceuticals necessitates expression in human or animal cell lines. These cell lines can be highly productive, with titers of overexpressed monoclonal antibodies (mAbs) exceeding 10g/L. The use of human and animal cell lines, however, introduces an inherent risk of contamination by undesirable adventitious viral agents. To address this concern, a holistic approach to virus safety is normally taken. This approach includes careful raw material selection, testing of cell lines, media components, and the use of appropriate viral inactivation and clearance protocols (1).

Virus clearance is typically achieved using at least two orthogonal methods: virus filtration and, low pH virus inactivation. ASTM E2888–12 states that this latter process should assure 5 log10 inactivation of non-defective C-type retroviruses when performed under the following parameters: hold temperature of ≥ 15 °C, hold time of ≥30 minutes, and pH of ≤3.6 throughout the course of the hold (2). For mAbs, low pH virus inactivation is typically performed directly after the Protein A capture chromatography step, as the elution pool is typically close to the required inactivation pH. In batch production, it is common for several Protein A eluates to be pooled and then subjected to the viral inactivation process, which essentially involves titration with acid followed by a hold time at the appropriate hold pH, typically pH3.5.

Continuous Protein A capture chromatography with a multicolumn chromatography system can generate multiple eluates each hour over multiple shifts or even days. To maximize the benefits of continuous capture chromatography, a semi-continuous or continuous solution for viral inactivation is required. 

Two basic approaches are viable for virus inactivation: a continuous plug flow reactor with in-line pH reduction, hold, and neutralization; and a multi-vessel continuously stirred-tank reactor (CSTR) with in-situ pH reduction, hold, and neutralization. While the CSTR approach generates a pause in the continuity in flow at the start of the process, advantages of this approach include:

• The approach is identical to the low pH inactivation method commonly employed by the industry and accepted by the regulatory authorities.

• Collection of elution pool enables homogenization of the protein and, therefore, eliminates any impact of protein concentration gradient on virus inactivation and simplifies virus clearance validation studies.

• The approach is transferable to other inactivation methods using solvent/detergent, etc., where the volume of the inactivation agent(s) is well characterized.

• CSTR serves as a suitable hold vessel for homogenization of solution conditions in order to prepare it for the subsequent purification steps. In this article, the authors describe the application of an automated, semicontinuous solution for low-pH viral inactivation for GMP production using the CSTR approach.

Semi-continuous virus inactivation system

The semi-contininous virus inactivation system (Cadence VI, Pall Biotech) Figure 1 consists of biocontainers for the acid and base, two 50-L Allegro mixers (Pall Biotech) with pH probes, and a central control system to direct fluid flow.

The two CSTRs are used alternately and asynchronously. A pool of eluates is collected in one of the mixers where acidification, hold, and neutralization takes place. While the inactivation process is occurring, the other CSTR is receiving new eluates. When the inactivation is complete in the first CSTR, the processed elution pool is transferred out, and the CSTR is ready to collect more eluates. Meanwhile, the inactivation cycle proceeds in the second CSTR.

A failure mode effects analysis (FMEA) highlighted two major risks in the system design: the accuracy of the pH probe over time and contamination of inactivated product with non-inactivated product.

Figure 1. Schematic of the Cadence V (Pall Biotech).

Evaluation of pH probes

The inactivation pH is a critical process parameter. To show that the inactivating pH has been achieved, the pH probes must remain precise and accurate throughout the continuous process (3). The reliability of a range of pH probes was evaluated over time in a simulated viral inactivation process. An automated test rig was designed to assess the probes, mimicking the process by transferring the probes between two human immunoglobulin G (IgG) 10g/L solutions (one in 20 mM acetate at pH 3.5 and the other at 20mM Tris at pH 8.0). Reference probes were used to monitor the pH of the two solutions over 48 hours.

Six re-useable probes exhibited minimal drift throughout the experiment; the accuracy of the pH probes at the beginning of the experiment, at 24 hours, and after 48 hours remained well within 0.1 pH units. Example data for a single probe are shown in Figure 2. As the re-useable probes exhibit minimal drift they were selected for use in the system. The probes are calibrated and autoclaved before sterile installation using an aseptic bellows design.

Minimizing contamination risk

The FMEA also identified the risk of untreated liquid coming into contact with inactivated product. This could be possible through “hanging drops” and “hold-ups”. To mitigate the risk of hanging drops, the CSTRs have ports for low-point loading to minimize splashing. Acid and base are added below the liquid level. To reduce the risk of foaming, a series of mixer experiments were performed. Various volumes of 10g/L IgG solution were mixed at increasing mixer speeds to determine the maximum mixer speed at a given volume that did not cause foaming. An algorithm, based on the data, can be used to adjust mixing speed based on volume.

A recirculation loop is used to ensure that the process fluid is homogeneous during the hold step and to prevent back mixing of acid and base. Valve blocks with a low hold-up design are employed.

Experiments were performed to determine if contamination due to hold ups and dead-legs occurred. Riboflavin was used as a visual test for hold-ups and hanging drops (4). The bacterium Brevundimonas diminuta (B. diminuta) was used to test the system for deadlegs. Experiments were also performed with the bacteriophage Phi6 as a mimic for enveloped mammalian viruses.

As a visual test, riboflavin was pumped into the system. No riboflavin could be detected above the original liquid level, and riboflavin could not be detected in the sampling port in the recirculation loop after recirculation. While a small amount of riboflavin was observed in folds in the biocontainer after draining and prior to the water rinse, no riboflavin was detected after refilling and subsequent emptying of the biocontainer.

B. diminuta(ATCC 19146) was selected as a surrogate for mammalian virus and offered increased method sensitivity (ability to detect one colony forming unit [CFU] in any given volume) and the ability to sample larger volumes by collecting bacteria via filtration before assaying. The test required a lower pH (pH 2.5) than is typically used as B. diminutais not killed at pH 3.5. Using this pH is consistent with the objective of the experiment; to assess if dead-legs could reduce the effectiveness of the pH treatment. 

Bacterial kill testing was performed over two continuous low pH cycles in a single mixer. After these two cycles were completed, sterile tryptic soy broth (TSB) was added to the mixer via the top port of the mixer biocontainer. This was aseptically collected, and 2 L (one-third of the volume) was analyzed by membrane filtration to assess the potential for carryover of viable organisms. Three samples were tested, one for each cycle and the wash-out of the system at the end with TSB. B. diminuta was not detected in any of the samples. A control experiment performed in the system omitting the low pH inactivation step showed no reduction in viable bacteria, 1.0 x 107 CFU per mL. This demonstrates that the system effectively kills B. diminuta with low pH and that no hanging drops or hold-ups reinfected the treated material. 

A mammalian virus analogue Phi6 was chosen to simulate the inactivation kinetics of enveloped mammalian viruses. Using bacteriophage allows a specific log reduction value (LRV) to be determined. The bacteriophage testing was performed using similar methods as the B. diminuta testing. Two low pH cycles were performed; the low pH inactivation for Phi6 was performed at pH 3.5 and the wash-out of the system was performed with buffer. Samples were analyzed for Phi6 using plaque assays with the host Pseudomonas syringiae, followed by overnight incubation at 37 °C. For Phi6, the data are shown in Table I.

No Phi6 were detected after low pH virus inactivation and no significant contamination of inactivated elution with non-inactivated elution occurred. Two cycles were performed in a single biocontainer indicating that there was no carryover of bacteriophage from cycle to cycle. In addition, a biocontainer wash after the second inactivation cycle showed no remaining active bacteriophage. A negative control (same pump and mixer speeds but without any low pH inactivation) was performed to confirm that the phage was not inactivated over time or due to shear forces from mixing and pumping. The results indicate that the titer was not changed without the low pH step and the system is effective for virus inactivation at low pH.

Performance over 24 hours

The system was operated for 24 hours to test performance without user input over prolonged periods. An IgG test solution containing 0.1x PBS titrated to pH 4 with acetic acid was used to mimic elutions from Protein A columns. To mimic for output from a continuous chromatography system, a pump set at a flow rate of 280 mL/ min was used to deliver 10 L per hour of the test solution cyclically (eight minutes pumping, four minutes no pumping).

An initial test titration revealed that to achieve pH values of 3.5 and 8, approximately 2.5% of the original volume of acetic acid and 5.2% of Tris had to be added, respectively. The titration strategy is shown in Table II.

The titration is performed in three steps. In step one, a bolus of titrant (approximately 50% of the expected amount) is added as a percent of the weight in the mixer. In step two, a smaller amount of titrant is added and the pH checked after a period of mixing. Rounds of addition and mixing are performed until the step two setpoint is reached. In step three, smaller amounts of titrant are added, mixed, and the pH checked. Further additions are made until the final step three checkpoint is reached. To avoid overshooting the pH set-point, normally these additions are relatively small.

After installation of a new manifold, a VI sequence was initiated, opening a flow-path for product to arrive in the CSTR. The pH probes were immersed for 10 minutes before a single set-point correction was performed to calibrate the system without pausing the process. No further calibrations were required during the 24-hour process. Samples were taken periodically to ensure that the pH of the system probes was accurate and represented the pH of samples in the mixers. It was found that the pH probes remained within 0.07 pH units of the reading from the externally calibrated probe and were thus within the desired accuracy of +/-0.1 pH units.

Data from the probes in the two CSTRs are shown in Figure 3. The pH of the incoming material was pH4. The material was titrated down to pH3.5, held at this pH for 60 minutes, and then titrated up to pH8. After pH8 was achieved, the CSTR was emptied and the probe exposed to air until new elutions covered it again and the cycle started over. The pH traces appear as expected, and no pH overshoots or process deviations were observed throughout the process. After completion of the 24 hours of operation, no apparent damage to the manifold and no leaks were observed.

Figure 3. A: Data showing the operation of the virus inactivation system over a 24-hour period via the two in-built pH probes. B: Focused view of a single virus inactivation cycle showing the short titration steps and accurate titration without overshoots.

Conclusion

The Cadence VI system (Pall Biotech) was designed to transfer batch virus inaction into an automated and continuous operation with the careful use of FMEA to mitigate risks. This resulted in a robust solution for the operation of fully automated, semi-continuous low pH virus inactivation, which may lead to a simple path to regulatory approval.

References

1. FDA, FDA: Guidance for Industry Q5A, Viral Safety Evaluation of Biotechnology Products Derived from Cell Lines of Human or Animal Origin(Rockville, MD, 1997). 
2. A.A. Shukla and H. Aranha, Pharm. Bioprocess.3(2) 127–138 (2015).
3. CMC Biotech Working Group, A-mAb: a Case Study in Bioprocess Development, Version 2.1, Oct. 30, 2009.
4. B. Kanegsberg and E. Kanegsberg, Handbook for Critical Cleaning: Applications, Processes, and Controls(CRC Press, Boca Raton, FL, 2011).

About the authors

Kyle Jones is R&D applications scientist II and Mark Schofield, PhD, is senior R&D manager; both are at Pall Biotech.