Downstream Processing: A Revalidation Study of Viral Clearance in the Purification of Monoclonal Antibody CB.Hep-1

ABSTRACT

This article revalidates the effectiveness of affinity chromatography, matrix sanitization, and storage procedures used in monoclonal antibody CB.Hep-1 purification to remove and inactivate viruses after process scale-up. The scale up of the CB.Hep-1 purification process demonstrated a similar removal factor for enveloped and nonenveloped viruses compared to the initial validation study. The HSV-1, HIV-1, and CPV viruses were sensitive to incubation with ethanol at 70% concentration (3.0–4.6 Logs). We found that 0.1N HCl is a robust chemical agent able to inactivate >6.13 Logs of nonenveloped high resistance viruses while ethanol at 20% concentration inactivated 3.7 Logs of enveloped viruses HSV-1 and HIV-1 but was unable to inactivate nonenveloped viruses HPV-1 and CPV.

Monoclonal antibodies (MAbs) are employed as immunoligands in the purification processes of biopharmaceutical products.^1,2 Virus transmission poses a high potential risk for patients who must be treated with these biopharmaceuticals, if MAbs come from human or animal sources.^3,4 It is necessary to validate the purification process capacity to remove and inactivate any potential viral contaminant.^5,6

Regardless of extreme virus controls, several instances of biological contamination have occurred. Research on virus contamination sources have shown that viruses can be introduced into the manufacturing process in different ways, illustrating the importance of viral clearance studies to guarantee the biopharmaceutical product safety.⁷

Validation of the purification method plays an essential role in establishing biological product safety, especially when there is a high risk for the source to be contaminated with known human pathogenic viruses. Since several contamination instances have occurred with agents whose presence was not known, validation also provides a high degree of confidence that these agents may also be removed.⁸

Acronym List

FDA defines validation as, "Establishing documented evidence, which provides a high degree of assurance that a specific process will consistently produce a product meeting its pre-determined specifications and quality attributes".⁹ The rationale is that if more effort is placed on validation at the beginning, then there will be less chance for failure during product life.¹⁰

Validation studies for purifications proceses involve the deliberate addition of a virus to one or more purification steps to measure the extent of its removal and inactivation capacity. It is not necessary to validate all purification steps, but only those that could contribute to virus removal or inactivation. To prevent the deliberate introduction of viruses into the manufacturing process, the validation studies should be done in a separate facility and in a scaled-down version of the manufacturing process. Validation at the small scale is an efficient way to perform viral clearance validation studies.¹¹

GENERAL PROCEDURE FOR VALIDATION

CB.Hep-1 MAb is a mouse IgG-2b, specific for HbsAg.¹² This MAb is used as an immunoligand in the antigen-purification process, which is one step in the manufacturing of the Hepatitis B vaccine for human use.^13,14 The main aim of our work was to investigate if the affinity chromatography used routinely in the purification of CB.Hep-1 shows the same virus removal factor after a scale-up process. We measured the virus inactivation factor of the column sanitization protocol using 70% ethanol and the matrix storage conditions in 20% ethanol. We also evaluated the column sanitization protocol with 0.1 N HCl to increase the inactivation factor for high resistance non-enveloped viruses.

The manufacturing process of MAb CB.Hep-1 is based primarily on Protein A–Sepharose affinity chromatography. In this work, the affinity chromatography's effectiveness in removing enveloped and nonenveloped viruses was revalidated after a scale-up of the purification process (scale-up factor = 6). Two earlier viral clearance validation studies of the MAb CB.Hep-1 purification process demonstrated that the unique step showing high removal capacity of model viruses was Protein A–Sepharose affinity chromatography.^15,16 For this reason, we only performed a revalidation study of this purification step.

Scale-down of a chromatographic process can generate different removal factors even when the same virus and resin are used.⁷ Several factors can be responsible for these differences; notably the buffer system, linear flowrate, bead height-to-diameter ratio, protein content, and temperature.¹⁷ We have reason to believe that the residence time is the most important parameter to scale-down affinity chromatography. The availability of small-scale commercial affinity chromatography columns facilitate this task. In this study, the scale-down process was specified so as to keep the residence time and the protein load-per-mL of matrix constant.

A major concern in validation clearance studies is determining which viruses should be used. Preference should be given to viruses with different shapes, genomes, sizes, envelopes, and resistance to physical and chemical agents. We selected our model viruses on the basis of recommendations from the literature and also because we wanted to use the same viruses used in the previous validation study. Detailed information about the selection appears in reference 15.

Figure 1. Flowsheet of monoclonal antibody CB.Hep-1 purification process. The study starts at the red arrow.

The number of viruses is another aspect that requires special attention. The number will depend on the level of characterization of the starting material and the production process. To demonstrate the viral removal factor of the MAb CB.Hep-1 scale-up purification process, HSV-1, HIV-1, HPV-2, and CPV were individually added to one tenth of the desalted material before the affinity chromatography step (Figure 1). To ensure the validity of this revalidation work, the model viruses used were the same viruses used during the first validation study (Table 1).¹⁸

Table 1. General characteristics of the model viruses used in the validation studies.

MATERIALS AND METHODS

We used the same MAb that was used in production. CB.Hep-1 MAb is an IgG-2b secreted by the hybridoma cell line 48/1/5/4. This hybridoma is a mouse–mouse hybridoma cell generated by the fusion of the HBsAg immunized BALB/c mouse lymphocytes and the heterohybridoma SP2/0-Ag14.¹²

Experiments were assessed using a scale-down version equivalent to 1% (1:100) of the MAb CB.Hep-1 manufacturing purification scale, which showed no significant differences in MAb yield (p = 0.0545) or specific activity (p = 0.0861). The recovery (p = 0.0299) and purity (p = 0.0327) showed significant differences with the Reference "22" validation study but all recovery and purity values were over the pre-determined specification limits (Figure 2). Therefore, the scale-down process performed within the limits established for the manufacturing purification scale and for the validation study.

Figure 2. Four graphs show that the yield, recovery, purity, and specific activity of the Mab CB.Hep-1 purification process are consistent in manufacturing scale and validation (laboratory) scale. Bars represent the average of the results of each parameter and the confidence limit. All experiments were performed in triplicate.

The model viruses used in this study, which are listed in Table 1, cover a wide range of physical–chemical and structural characteristics of viruses.¹⁸ Sources for the four viruses are listed in Table 2. Cell lines employed in the virus titration were: African green monkey kidney cell line (Vero), guinea pig fibroblast cell line (LFBC) and the HIV negative human T cell line (MT4). The source of these cell lines is listed in Table 2. The log TCID50/mL virus titers in the starting stocks were: 8.4 for HSV-1; 5.8 for HIV-1; 6.5 for HPV-2; and 7.5 for CPV. All the model viruses were purified by an ultracentrifugation method and then filtered under sterile conditions in separate laboratories and at different times to ensure that each virus preparation was pure.

Table 2. Source of the cell line used for titration and detection of model viruses in both validation and revalidation studies

Protein A–Sepharose affinity chromatography viral removal

This revalidation study was carried out following the same principles as the viral validation study.¹⁵ It was performed in a separate laboratory using the same affinity-chromatography residence time and protein concentration in the column's applied material as in production. The purification process flow is illustrated in Figure 1. The starting material was individually loaded with each model virus and spiked into the Protein A–Sepharose affinity column (three independent replicas for each virus). The IgG adsorption buffer was PBS (pH 8.0) and 0.1 M citric acid (pH 3.0) was used as the elution buffer. All experiments were conducted at 4 °C. The affinity column used was an Amersham-Bioscences XK26/30 loaded with 54.2 mL of matrix and operated at 100 cm/h of linear flowrate. The protein absorbance was registered using a 280 nm filter and the applied IgG per run was 90% of the dynamic IgG binding capacity of the matrix.

Samples for virus detection were taken from each elution fraction (not from the pool of elution fractions). These fractions were then dialyzed with PBS to raise the pH. Because viruses were detected in the elution fractions, we believe viral clearance occurred primarily as a result of virus removal, but inactivation could not be completely avoided, because, as we have demonstrated, enveloped viruses are easily inactivated at low pH.

Quantification of MAb CB.Hep-1 by ELISA

A Costar polystyrene microplate (Corning Life Sciences) was coated with 10 μg per well of rHBsAg in 0.1 M NaHCO₃ buffer for 20 min at 50 °C. After this step, samples were added to the plate in 0.05% Tween 20 in PBS and incubated for 1 h at 37 °C. Several washes with 0.05% Tween 20/PBS were done. Then the plate was incubated for 1 h at 37 °C with a horseradish peroxidase conjugate (Sigma Chemical). The reaction was then revealed using 100 μL/well of 0.05% O-phenylenediamine and 0.015% H₂O₂ in citrate buffer (pH 5.0), and stopped with 50 μL/well of 1.25M H₂SO₄. The absorbance was measured in a Labsystems Multiskan enzyme-linked immunosorbent assay (ELISA) reader using a 492-nm filter. The range of the calibration curve was from 3.13 to 50 ng/mL. The inner standard MAb used was IgG2B070305, supplied by the Quality Control Department of Center for Genetic Engineering and Biotechnology, Havana, Cuba.

Determination of MAb CB.Hep-1 purity by SDS-PAGE

Samples were analyzed by electrophoresis on 12.5% sodium dodecyl sulphate poliacrylamide gel electrophoresis (SDS-PAGE) gels as described by Laemmli.¹⁹ Separated proteins were stained with Coomassie blue R-250 and then analyzed by densitometry. Percentage of purity was measured by Bio-RAD Version 1.4.1 Build 446 Molecular Analyst software. All samples (20 μg) were applied to the gel under reduction conditions.

Virus inactivation during affinity matrix sanitization with 70% ethanol

Nine mL of the Protein A–Sepharose matrix were previously washed and equilibrated with a fivefold column volume of 70% ethanol/30% purified water. The model viruses (HSV-1, HIV-1, HPV-2, and CPV) were individually diluted (1:10 v/v) and added to the matrix. The working temperature was 4 °C and supernatants were collected after each exposure time (0 min and 12 h) to be dialyzed against PBS (pH 7.2) to allow the virus titration. The control was a similar amount of virus added to the Protein A–Sepharose matrix previously equilibrated with PBS (pH 7.2) and kept at 4 °C during the whole experiment.

Nonenveloped virus inactivation during affinity matrix sanitization with 0.1 N HCl

The Protein A–Sepharose matrix was previously washed and equilibrated with a fivefold column volume of 0.1 N HCl (pH 1.0). Nonenveloped model viruses (HPV-2 and CPV) were diluted 1:10 (v/v) and added to the matrix. The working temperature was 4 °C and samples were collected after the following exposure times: 0 min, 3, 4, 5, 8, and 10 h to be dialyzed against PBS (pH 7.2) to allow the virus titration. The control was a similar amount of virus added to the affinity matrix previously equilibrated with PBS (pH 7.2) and kept at 4 °C during the whole experiment.

Virus inactivation during affinity matrix storage with 20% ethanol

Nine mL of the Protein A–Sepharose matrix previously washed and equilibrated with a fivefold column volume of 20% ethanol/80% purified water. The model viruses (HSV-1, HPV-2, HIV-1, CPV) were individually diluted (1:10) were added to the matrix. The working temperature was again 4 °C and samples were collected after each exposure time. For HSV-1 and HIV-1: 0 min, 15 min, 1 h, and 2 h; for HPV-2 and CPV: 0, 15, and 30 min, and 1, 24 h, 48 h, and 72 h. All samples were dialyzed against PBS (pH 7.2) to allow the virus titration. The control was a similar amount of virus added to the affinity matrix previously equilibrated with PBS (pH 7.2) at 4 °C.

TITRATION

Human herpes simplex type 1 and human poliovirus type 2 titration

The HSV-1 and HVP-2 titers were determined by the inoculation of test solutions into Vero cell cultures (Table 2) and calculated using the Reed Müench method.²¹ Briefly, the assay consisted of inoculating 50 μL of viral sample in 150 μL of Gibco minimum essential medium (MEM) supplemented with 2% of Hyclone FCS containing Vero confluent cells. Ten serial dilutions were performed across the plates. Plates were maintained at 37 °C under 5–6% CO₂ atmosphere. On the fifth day, cultures were carefully observed. End points were taken as the last dilution given cytopathic effect, with virus titer expressed as log₁₀TCID50 mL^-1. The same virus preparation was used as control of the experiment, and it was storage-aliquoted at –70 °C until thawing immediately prior to the titration assay. The virus titration was considered satisfactory when the difference between the expected and the true titer of the HVS-1 or HPV-2 used as control was less than 1 Log. Each sample was titrated in triplicate with eight determinations per each serial dilution.

Human immunodeficiency virus type 1 titration

The HIV-1 titers were determined by the inoculation of samples into MT4 cell culture (Table 2) and calculated using the Reed Müench method.²⁰ Fifty μL of viral samples were inoculated in 150 μL of Gibco RPMI 1640 supplemented with 2% of Hyclone FCS containing MT4 confluent cells. Ten serial dilutions were performed across the plates. Plates were maintained at 37 °C under 5–6% CO₂ atmosphere. On the seventh day the cultures were carefully observed and the virus titer was expressed as log₁₀TCID50 mL^-1 of the last dilution with cytopathic effect. The same virus preparation was used as control of the titration assay. The virus titration was considered satisfactory when the difference between the expected and the true titer of the HIV-1 used as control was less than 1 Log. Each sample was titrated in triplicate with eight determinations per each dilution.

Canine parvovirus titration

The titers of CPV were determined according to the Reed Müench method by inoculating of test solutions into LFBC cell cultures (Table 2).²⁰ The viral samples were added to 150 μL of MEM supplemented with 2% of FCS containing LFBC confluent cells and maintained at 37 °C under 5–6% CO₂ atmosphere. On the fifth day, the culture was carefully observed. The last dilution with virus cytopathic effect was taken to estimate the virus titer. The same CPV preparation was used as a control of the experiment and the titration assay also considered satisfactory when the difference between the expected and the true titer of the CPV used as control was <1 Log. Each sample was also titrated in triplicate with eight determinations per dilution.

MATHEMATICAL ANALYSIS

The Kruscall-Wallis test was executed to compare the yield, recovery, purity, and specific activity results between the scale-down and the manufacturing scale of the MAb CB.Hep-1 purification process. The Mann-Whitney (Wilcoxon) W test was used to compare medians of two samples, sorting the data from smallest to largest, and comparing the average ranks of two samples in the combine data. The program used was the STATGRAPHICS plus 5.0, 1994–2000 (Statistical Group Graphic Corporation).

Calculation of removal or inactivation factor (RF)

Each virus removal factor was calculated individually by White-Grun-Sur-Sito method as in Equation [1], below:²¹

RF is the log of Equation [1].

RESULTS AND DISCUSSION

When the two studies were compared, the results demonstrated no significant differences in the removal factors of these model viruses. These results corroborate that the scale-up of the MAb CB.Hep-1 purification process does not modify the capacity of Protein A–Sepharose to remove the HSV-1 (p = 0.0808), HIV-1 (p = 0.6625), HPV-2 (p = 0.1904), and CPV (p = 0.0765). As expected, affinity chromatography showed higher removal capacity for enveloped and large viruses (Figure 3). The percentage reductions were 97.55%±2.42 (HSV-1) and 90.29%±10.46 (HIV-1). Small and nonenveloped viruses were removed in lower percentages: 47.42%±1.11 (HPV-2) and 62.79%±5.24 (Figure 3), which appears to occur because the small size of these viruses affects their interaction with and penetration into the matrix, thus retarding the elution of virus particles.

Figure 3. Comparison of two validation studies of virus removal factor of the Protein A–Sepharose affinity chromatography. Bars show percentage of the applied virus removed by chromatography. All experiments were done in triplicate.

Affinity column (20%) carryover studies showed virus presence in the product after subsequent runs.²² Thus, matrix sanitization and storage protocol effectiveness demonstration is also mandatory for viral clearance studies. Robust virus clearance studies are defined as those that have been shown to work accurately under a variety of conditions such as pH or ionic strength of column buffers. Virus inactivation may be achieved by a number of physical (heat, radiation, and sonication) or chemical (detergents, alcohol, solvents, acids, bases, glutaraldehyde, and B-propiolactone) methods.^3,23 It is also important to consider that enveloped RNA and DNA viruses have low resistance to physical–chemical agents, because solvents like ethanol destroy their envelopes, which are composed of proteins and lipids from the cells. Conversely, the lack of an envelope makes a virus quite resistant to these agents.

Our results also confirmed the expected results for the enveloped viruses and for the HPV-2. The viral inactivation study performed in this work demonstrated 3.0–4.6 Logs of inactivation for enveloped viruses HVS-1 and HIV-1, respectively, in 70% ethanol and greater than 3.7 Logs in just 2 h in 20% ethanol for both viruses (Figure 4). It is consistent that the nonenveloped virus HPV-2 retains its infection capacity in both ethanol concentrations.

Figure 4. Virus inactivation factors of the Protein A–Sepharose affinity column sanitization and storage protocols. Bars represent the average and the confidence limit of the inactivation factor. All experiments were carried out in triplicate.

The exception to our theory was that the 70% ethanol was able to inactivate 3.97 Logs of the high-resistance CPV, thereby demonstrating that the lack of an envelope is not a necessary for these kinds of viruses to be able to resist the chemical agents (organic solvents). We hypothesize that the proteins involved in the virus infectivity mechanism of the CPV suffer some kind of chemical modification that blocks its capacity to infect the host cells. The 20% ethanol was totally inefficient for inactivating (<1.3 Logs) the CPV in 72 h (Figure 4).

To increase the inactivation factor of the whole sanitization procedure, the resistance of nonenveloped viruses HPV-2 and CPV was also studied in 0.1 N HCl (pH 1.0). The isoelectric pH of the purified poliovirus and CPV ranges from 5.5 to 7.0.²⁴ This attribute will be drastically changed under the presence of the 0.1 N HCl allowing, in theory, a high virus inactivation factor. Results demonstrated this. HPV-2 was inactivated 6.13 Logs in 7 h and CPV was inactivated 6.60 Logs in just 1 h. Therefore, a robust sanitization protocol for the affinity chromatography step of a MAb CB.Hep-1 purification process should involve the combination of 70% ethanol and 0.1 N HCl in the studied exposure time.

CONCLUSIONS

A scale down of the MAb CB.Hep-1 purification process demonstrated removal factors of enveloped and nonenveloped model viruses similar to the first validation study. HSV-1, HIV-1, and CPV are sensitive to incubation with 70% ethanol (from 3.0 to 4.6 Logs). We find that 0.1 N HCl provides robust inactivation of more than 6.13 Logs of nonenveloped high resistance viruses. Ethanol at 20% concentration inactivated 3.7 Logs of enveloped viruses HSV-1 and HIV-1 but was totally ineffective in inactivating the nonenveloped viruses HPV-1 and CPV.

ACKNOWLEDGMENTS

The authors kindly thank the Cuban National Center for Animal Breeding for producing and supplying the specific pathogen-free ascitic fluid used in this work.

Rodolfo Valdés is head of the monoclonal antibody production department at the Center for Genetic Engineering and Biotechnology (CGEB), Havana, Cuba, 537.2716022, ext. 7101, rodolfo.valdes@cigb.edu.cu

María del Rosario Alemán is the head of the cell culture laboratory of the monoclonal antibody production department at CGEB.

Andrés Tamayo, Sigifredo Padilla, Lamay Dorta, and Biunayki Reyes are researchers at the monoclonal antibody production department at CGEB.

Enrique Noa, Marta Dubed, and Giselle Alvarez are researchers at the National Reference Center for AIDS Research, Havana, Cuba.

Déborah Geada and Leonor Navea is a researcher at the Research Institute of Tobacco, Department of Industrial Development, Havana, Cuba.

Tatiana González is an engineer at the Process Control Department of CGEB. Eutimio Gustavo Fernández is a researcher at Inorganic Chemistry Department, Center for Engineering and Chemical Researches, Havana, Cuba.

REFERENCES

1. Quiñones Y, Agraz A, Silva A, Padrón G, Mella C, Díaz R, et al. High purity recombinant human alpha-2 interferon free from oligomeric forms in E. Coli. Highlights Modern Biochem 1989; 2:1237–1246.

2. Stum DC, Thienpont M, Collen D. Isolation y characterization of single chain urokinase (prourokinase) and urokinasa-inhibitor complex. J. Biological Chemistry 1986; 261:1267–1273.

3. Harbour C, Woodhouse G. Viral contamination of monoclonal antibody preparations: potential problems and possible solutions. Cytotechnology 1990; 4:3–12.

4. Hart CA. Transmissible spongioform encephalopathies. J Med Microbiol. 1995; 42:153–155.

5. Borovec S, Broumis C, Adcock W, Fang R, Uren E. Inactivation kinetics of model and relevant blood-borne viruses by treatment with sodium hydroxide and heat. Biologicals 1998; 26, 3:237–244.

6. Committee for Proprietary Medicinal Products. EEC regulatory document, note for guidance: Validation of viral removal and inactivation procedure. Biological, 1991; 19:247–25.

7. Darling JA, Spaltro JJ. Process validation for virus removal. Biopharm 1996; 10:42–50.

8. Minor P. Adventitious viral agents in biological product. Develop. Biol. Standard 1989; 70:173–179.

9. FDA. Guideline on general principles of process validation. CDER. Bethesda Md: 1978 May.

10. Juran JM, Gryna MF, editors Juran's quality control handbook, 4th ed. New York: McGraw-Hill; 1988.

11. Sofer G. Validation: ensuring the accuracy of scaled-down chromatography method. Biopharm 1998; 10:51–54.

12. Fontirrochi G, Dueñas M, Fernández E, Fuentes P, Pérez M, Mainet D, et al., A mouse hybridoma cell line secreting IgG and IgM with specificity for the Hepatitis B surface antigen virus requirements for biological substances. Biotecnología Aplicada 1993; 10:25–30.

13. Hardy E, Martínez E, Diago D, Díaz R, González D, Herrera L. Large scale production of recombinant Hepatitis B surface antigen from Pichia pastoris. J. Biotechnology 2000; 17:155–157.

14. Fernández ME, Díaz T, Galván A, Valdés R, González E, Ayala M, et al. Antigen recognition characteristics and comparative performance in immunoaffinity purification of two monoclonal antibodies specific for the Hepatitis B virus surface antigen. J Biotechnol. 1997; 56:69–80.

15. Valdés R, Ibarra N, Ruibal I, Bederraín A, Noa E, Herrera N, et al. Chromatographic removal combined with heat, acid and chaotropic inactivation of four model viruses. J. Biotechnology 2002; 96:251–258.

16. Valdés R, Díaz T, Nieto A, García C, Pérez M, García J, Quiñones Y. Sendai virus removal and inactivation during monoclonal antibody purification. Biotecnología Aplicada 2005; 12 (2):115–119.

17. Kelly DB, Jeening P, Wright R, Briaco C. Demonstrating the process robustness for chromatographic purification of a recombinant protein. Biopharm 1997; 10:36–45.

18. Berthold W, Walter J, Werz W. Experimental approaches to guarantee minimal risk of potential virus in purified monoclonal antibodies. Cytotechnology 1992; 9:189–201.

19. Laemmli. UK. Cleavage of structural proteins during the assembly by bacteriophage T 4. Nature 1970; 270:680–685.

20. Reed. LJ, Müench H. A simple method of estimating fifty per cent endpoints. Amer. J. Hyg. 1938; 27:493–487.

21. White EM, Grun BJ, Sur C-S, Sito FA. Process validation for virus removal and inactivation. BioPharm 1991; 4(5):30–33.

22. Willkommen H. Safety issue with recombinant technologies, assessment of clinical trial material. Bioprocess Int. Prague, 2006. Feb 21.

23. Cameron R, Davies J, Adcock W, Mac Gregor A, Bargord RJ, Cossart Y, Harcour C. The removal of model virus, poliovirus type 1 and canine parvovirus, during the purification of human albumin using ion-exchange chromatography procedures. Biologicals 1997; 25:391–401.