ANALYTICAL SCIENCES APRIL 2014, VOL The Japan Society for Analytical Chemistry

Transcription

1 ANALYTICAL SCIENCES APRIL 2014, VOL The Japan Society for Analytical Chemistry Application of Microchip Electrophoresis Sodium Dodecyl Sulfate for the Evaluation of Change of Degradation Species of Therapeutic Antibodies in Stability Testing Yuki YAGI,* Kazuaki KAKEHI,** Takao HAYAKAWA,** and Shigeo SUZUKI** * Kyowa Hakko Kirin Co., Ltd., Takasaki, Gunma , Japan ** Faculty of Pharmaceutical Sciences, Kinki University, Kowakae, Higashi-Osaka , Japan We evaluated the performance of a commercial microchip electrophoresis instrument (LabChip GXII) for the evaluation of change of degradation species of therapeutic antibodies in stability testing. This system requires a sample volume of only 5 μg, and indicates fine resolution of size variant species such as light chain, heavy chain, non-glycosylated heavy chain and various degradation species. Precision and accuracy were high; the intermediate precision of 18 determinations was only 2.1% or less as RSD and recoveries ranged from 97.8 to 103.0% for major species as heavy chain, light chain and intact molecule of a therapeutic antibody. The applicability of this method was demonstrated by applying the method for the analysis of heat-degraded products of three pharmaceutical antibodies. Though some fragment peaks commonly appeared and increased according to temperature regardless of the source of preparations, one of them indicated specific peaks implying the cleavage of the peptide chain of the heavy chain. We also compared the performance of the method with those using conventional capillary-based SDS electrophoresis. Although the absolute purity values expressed as peak area % were different for the two methods, probably due to the difference in the detection methods, similar quality profiles were obtained within 40 s by microchip-based SDS electrophoresis. In addition, the degradation manner of three marketed antibodies depending on temperature was almost the same for the two methods. At the first stage in the development of manufacturing antibody pharmaceuticals, various factors including cell selection, cell cultivation, and formulation development should be evaluated using limited sample amounts. The stability testing using microchip-based electrophoresis seems suitable for these purposes. Keywords Microchip electrophoresis, capillary electrophoresis, antibody (Received December 17, 2013; Accepted February 20, 2014; Published April 10, 2014) Introduction Monoclonal antibodies are important therapeutic agents because of their high specificity and low toxicity. More than 20 therapeutic antibodies have been approved for the treatment of several severe diseases. 1 The production of these antibodies requires well controlled manufacturing processes because impurities or degradation could occur depending on manufacturing and storage processes, and these impurities or degradation often cause toxicity and reduction of efficacy. These heterogeneities include aggregation, fragmentation, deamidation, oxidation, N- and C-terminal differentiations Several analytical methods have been developed for these purposes, for example, SDS polyacrylamide gel electrophoresis for the determination of the purity of proteins, size exclusion chromatography for aggregation, ion exchange chromatography or hydrophobic interaction chromatography for the detection of deamidation and oxidation variants, and Ellman s assay for quantification of free thiol groups. 12,13 Capillary-based SDS electrophoresis (CE-SDS) is one version To whom correspondence should be addressed. yuuki.yagi@kyowa-kirin.co.jp of the capillary gel electrophoresis (CGE) using soluble linear polymers. Agarose and cross-linked polyacrylamide were often used as sieving matrices in the 1980s, 14,15 but the life time and reproducibility were poor. Availability of soluble linear polymers improved reproducibility and robustness of CGE. SDS is bound to all the proteins at a constant weight ratio such that the SDS-protein complexes have approximately equivalent charge densities. The constant mass-to-charge property of the SDS-bound proteins allows separation according to differences in molecular weight. CE-SDS has been recognized as an important tool for the characterization of therapeutic antibodies In the commonly used CE-SDS method, proteins are separated based on the size differences in a capillary filled with soluble linear polymer solution, and the separated components are detected by UV or fluorometric detectors. The CE-SDS method is superior to the classical SDS polyacrylamide gel electrophoresis in resolution, reproducibility, and quantitativity. But, it takes about 20 min for separation and requires additional time for sample preparation, rinsing capillary tubes and loading the polymer. In addition, problems can arise due to injection failure, appearance of spike noise, and drift and shift of the baseline. Therefore, it requires a long time to apply CE-SDS for the analysis of antibody fragmentation. Microchip devices were adopted for use in electrophoresis

2 484 ANALYTICAL SCIENCES APRIL 2014, VOL. 30 from the 1990s. 22 Microchip electrophoresis has the potential for high throughput, and has been successfully used in nucleic acid analysis. Applications of microchip electrophoresis on protein separation have been reported by several research groups, and several kinds of commercial instruments have been launched The LabChip system LCGXII (Perkin Elmer, Inc.) is a commercial instrument that has features enabling automated analysis of a sample within 40 s, and can handle 96 or 384 samples at once. Moreover, with this system, the separated proteins are automatically stained and destained on a chip. The evaluation of non-glycosylated heavy chain, glycan type assay and dimer screening assay using the LabChip system was implemented by Chen et al. 27,28 Their studies mainly focused on the quality evaluation of the antibody such as for clone selection or cell culture optimization. There are some cases in which the development of the antibody stops due to stability defects even if the quality of the antibody immediately after manufacture is good. We applied here the method for heat degradation analysis of therapeutic antibodies using microchip-based SDS electrophoresis (ME-SDS). Application for stability testing and comparison of the separation with CE-SDS, focusing on the change in the degradation of therapeutic antibodies by heat stressing, should provide useful information for manufacturing therapeutic antibodies. Experimantal Materials The following materials were used in this study: HT protein express kits with chips and reagents (Perkin Elmer, Inc., MA), SDS gel buffer, bare fused-silica capillary and 5-TAMRA.SE (Beckman Coulter, Inc., Fullerton, CA), NAP-5 columns (GE Healthcare Life Sciences, Piscataway, NJ), SDS, disodium hydrogenphosphate, citric acid, sodium bicarbonate (Wako Pure Chemical Industries, Tokyo, Japan), 2-mercaptoethanol (BME, Nacalai Tesque, Tokyo, Japan), dithiothreitol (DTT, Life Technologies, Carlsbad, CA), DMSO and iodoacetamide (Sigma-Aldrich, St. Louis, MO). Our research group also obtained three therapeutic antibodies, mab A, B and C, in commercially available form. ME-SDS The ME-SDS analysis was performed using an LCGXII system (Perkin Elmer, Inc.), which enables automation in the sample loading, staining, separation, destaining and detection. Prepared samples were vacuumed from a microtiter plate and electrokinetically loaded and injected into a separation channel measuring 14 mm in length and 31 μm in width. The separation buffer was prepared by mixing the HT protein express gel matrix and a fluorescent dye. The LIF detector used a red laser diode with excitation and emission wavelengths of 635 and 700 nm, respectively. Sample solutions were prepared by the method described by Chen. 27 Briefly, 5 μl antibody samples at 1 mg/ml were mixed with 35 μl denaturation solution. The denaturation solution was prepared by mixing with 700 μl HT protein express sample buffer with either 24.5 μl BME for the reducing assay, or 70 μl of 250 mm iodoacetamide for the non-reducing assay. The samples were incubated at 70 C for 15 min. Then, 70 μl of water was added to each sample. Separation was performed by HT Antibody 200 mode on the instrument. CE-SDS The CE-SDS analysis was performed using a Proteomelab PA800 system with LIF detector using 488 nm of an argon laser as a light source and an emission band-pass filter of 560 nm (Beckman Coulter, Inc.) with a fused silica capillary (Beckman Coulter, Inc.) of 50 μm i.d., a total length of 30 cm and effective length of 20 cm. Samples were prepared according to the method described by Hunt et al. 17 Briefly, antibody samples were equilibrated with a buffer by passing through a NAP-5 column previously equilibrated with 0.1 M sodium bicarbonate, ph 8.3. A 10-μL portion of 5-TAMRA.SE dissolved in DMSO at a concentration of 0.3 mm was added to 190 μl of antibody solution. Then, the solution was incubated 30 min at 30 C. Excess dye was removed by NAP-5 and the buffer exchanged into 85 mm citrate phosphate buffer (ph 6.5). The labeled sample was mixed with either SDS solution containing DTT for the reducing assay or SDS solution containing iodoacetamide for the non-reducing assay, and incubated at 65 C for 5 or 10 min, respectively. The SDS gel buffer was loaded into the capillary at 15 kv for 10 min from the outlet side. A sample solution was injected electrokinetically at 10 kv for 10 s. The separation was conducted in the negative polarity mode at 15 kv for 20 min. Results and Discussion Method qualification for ME-SDS A ME-SDS method should fulfill the requirement for quantification and qualification of therapeutic antibodies as well as offer the capability for specific detection of minute amounts of degradation species existing in a large amount of native IgG molecules. At first, we implemented various parameters including reproducibility, repeatability, precision, accuracy, linearity and limit of quantitation using mab A as a standard sample. IgG antibodies are molecules about 150 kda composed of two light chains (LC, about 25 kda) and two heavy chains (HC, about 50 kda), and they are linked through disulfide bonds. Under non-reducing conditions, a large peak for mab A appeared at 36.0 s (Fig. 1A). An enlarged electropherogram (Fig. 1C) indicates five small peaks at 22.3, 25.7, 29.7, 32.6 and 34.4 s. From the molecular weight assigned by a molecular weight marker, they are estimated to correspond to an LC (30.1 kda), an HC (53.4 kda), a dimer of LC and HC (89.6 kda), a dimer of two HCs (115.7 kda), and a trimer of one LC and two HCs (141.4 kda). All peaks were estimated to be derived from IgG. Meanwhile, under reducing conditions, IgG molecules were split into two LCs and two HCs, and they appeared at around 22.4 and 27.2 s, respectively, as shown in Fig. 1B. In contrast to Fig. 1C, no minor peaks appeared in the expanded view (Fig. 1D) of the separation of reduced mab A. This result indicates the mab A sample does not contain cleavage products of IgG peptide chains. A small peak just in the front of an HC was assignable to be non-glycosylated HC. The fluorescent dye used in ME-SDS forms a complex with SDS and non-covalently binds to the hydrophobic regions of proteins. Thus, the signal intensity may depend on the hydrophobicity of peptides and proteins, but it is not critical for the evaluation of the purity alteration. Validation of ME-SDS was performed according to ICH guidelines, and the obtained results are listed in Table 1. Injection reproducibility, repeatability and intermediate precision were carried out as given in the footnote of Table 1, and the relative standard deviations of these parameters showed acceptable levels for quality control. These values compare

3 ANALYTICAL SCIENCES APRIL 2014, VOL Fig. 1 ME-SDS analysis of mab A prepared in non-reducing conditions (A and C) and reducing conditions (B and D). (C) and (D) are expanded views of (A) and (B), respectively. LM, lower marker; SP, system peak; LC, light chain; HC, heavy chain; LMWS, low molecular weight species; MMWS, middle molecular weight species; HMWS, high molecular weight species; NGHC, non-glycosylated heavy chain. Table 1 Reproducibility, repeatability, intermediate precision, accuracy and linearity of ME-SDS Parameter Non-reducing Reducing IgG, % LC, % HC, % Injection reproducibility a Repeatability (n = 6) b Intermediate precision (n = 18) c Accuracy Linearity Mean RSD, % Mean RSD, % Mean RSD, % Recovery, % d r e a. One preparation of non-reduced and reduced antibody samples were analyzed six times by iterative injections. b. Six individual preparations of non-reduced and reduced antibody samples were analyzed by single injection for each sample. c. Single analysis of six individual preparations of non-reduced and reduced antibody for three days. d. Determined by triplicated injection at concentrations of 80, 100 and 120% of mab preparations (1 mg/ml). e. Estimated from the calibration plots for two analysis each of 80, 90, 100, 110 and 120% of mab (1 mg/ml). favorably with that published for CE-SDS analysis. 19 An evaluation of accuracy and linearity over the range of % of the target concentration (1 mg/ml) resulted in a good recovery rate and correlation coefficient, respectively. To evaluate the limit of quantitation of the assay, BSA was added to antibody samples at concentrations of, 0.4,, 0.8 and 1.0%. The recoveries of BSA in all preparations were more than 80%. It was thus confirmed that the ME-SDS analysis can detect minute impurities. These results demonstrated that the ME-SDS is highly reliable and suitable for quality control in the manufacturing process. Application for stability testing Three mabs were incubated at 25 or 40 C for one month and the degraded products were analyzed by ME-SDS in non-reducing and reducing conditions. Though the mabs indicated analogous degradation profiles in ME-SDS, mab B gave additional peaks in front of HC and low molecular weight area under reducing conditions. We describe the results obtained for mab B in this section. Representative electropherograms of non-reduced and reduced mab B are shown in Fig. 2. Some fragment peaks (indicated by triangle marks) increased depending on temperature. The electropherogram for mab B showed the increase of peak a and b in non-reducing conditions and peak b and c in reduced conditions that was not detected for mab A and mab C (data not shown). The molecular weight of peak a (about 142 kda) and peak c (about 40 kda) was slightly smaller than IgG and HC, respectively, and peaks b and b had the same molecular mass (about 14 kda). This result strongly suggests that the cleavage of the HC chain occurred under thermal degradation conditions. When manufacturers found the cleavage of LC or HC chains in stress testing, they could change the amino acid sequence to protect from self-autolysis or could prolong the products by revising

4 486 ANALYTICAL SCIENCES APRIL 2014, VOL. 30 Fig. 2 Expanded electoropherograms of heat degradation of mab B. Unstressed, 25 and 40 under non-reducig conditions (A) and reducing conditions (B). Triangles show the minor peak, which increased by heat degradation. The structure of each peak is presumed from the molecular weight. Table 2 Analysis of degradation species by ME-SDS Condition Sample LMWS, % IgG, % (purity) Non-reducing Unstressed Condition Sample LMWS, % LC, % MMWS, % NGHC, % HC, % HMWS, % LC + HC, % (purity) Unstressed Reducing LMWS, low molecular weight species; MMWS, middle molecular weight species; HMWS, high molecular weight species; NGHC, non-glycosylated heavy chain. formulation buffers. This indicates ME-SDS would be useful in the analysis of thermal degradation profiles. Degradation species included in unstressed and heat-degraded mab B samples are shown in Table 2. With an increase in temperature, the intensity of minor peaks of the degraded products apparently increased except for nonglycosylated HC indicated as NGHC. These results showed that the assay could detect changes in preparation and storage processes, which may be helpful in the quality control of therapeutic antibodies. Comparison with the CE-SDS As described in the above sections, microfluidic separation devices have the potential to analyze therapeutic antibodies with high efficiency, and the 14 mm separation channel shortens analysis time to within 40 s. However, the short separation channel could cause defects in the resolution. To evaluate the performance of ME-SDS, thermally degraded samples of mab B were analyzed both by ME-SDS and by conventional CE-SDS, and their separation profiles were compared. Samples of mab B unstressed and stored at 40 were prepared under non-reducing conditions and analyzed by both ME-SDS and conventional CE-SDS. As shown in Fig. 3, similar separation profiles were obtained for both methods. Moreover, the intact IgG and peak 7 corresponding to nonglycosylated IgG were not separated in CE-SDS, and high resolution was obtained in ME-SDS. In reducing conditions, minor peaks including degradation species of HC appeared prior to non-glycosylated HC and some high molecular weight species were detected by both methods. Moreover, these two modes indicate similar electropherograms despite the difference in the detection methods. As expected, mab A and C exhibited similar separation profiles with both methods. Figure 4 shows the quantitative results obtained for the nonreduced and reduced samples prepared by applying heat stress

5 ANALYTICAL SCIENCES APRIL 2014, VOL Fig. 3 Comparison of ME-SDS and conventional CE-SDS. mab B stored at 40 was analyzed by the two methods. ME-SDS under non-reducing conditions (A) and reducing conditions (C). CE-SDS analysis under reducing conditions (B) and reducing conditions (D). Peak assignments were the same as those for Fig. 2. Fig. 4 Purity alteration of mab A, mab B and mab C. ME-SDS analysis in non-reducing conditions (A) and reducing conditions (C). CE-SDS analysis in reducing conditions (B) and reducing conditions (D). on three mabs. The observed purity percent values differed between the two methods. The difference is thought to be partly due to the detection responses of fragment peaks in the two methods. However, it is apparent that the degradation manner of the antibodies was comparably estimated by the two methods. Conclusions ME-SDS assay for therapeutic antibodies has the potential to separate degradation species with high reproducibility, precision and accuracy. The resolution of both of non-reduced and reduced antibodies were completed within 40 s under simplified

6 488 ANALYTICAL SCIENCES APRIL 2014, VOL. 30 operation procedures. The method is free from problems such as injection failure, the occurrence of noise peaks and baseline shift often observed in CE-SDS. ME-SDS requires antibody samples of only 5 μg in volume, which enables stability testing at the early stage of studies in the production of antibody pharmaceuticals, including cell selection, cell cultivation, and formulation development in small scale. We thus expect the ME-SDS could become an alternative method to the commonly used CE-SDS methods. Acknowledgements This work was supported by a Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (JSPS), Japan, References 1. J. G. Elvin, R. G. Couston, and C. F. van der Walle, Int. J. Pharm., 2013, 440, A. Eon-Duval, H. Broly, and R. Gleixner, Biotechnol. Prog., 2012, 28, S. Barandun, P. Kistler, F. Jeunet, and H. Isliker, Vox Sang., 1962, 7, E. F. Ellis and C. S. Henney, J. Allergy, 1969, 43, A. J. Cordoba, B. J. Shyong, D. Breen, and R. J. Harris, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2005, 818, M. Page, C. Ling, P. Dilger, M. Bentley, T. Forsey, C. Longstaff, and R. Thorpe, Vox Sang., 1995, 69, Y. D. Liu, A. M. Goetze, R. B. Bass, and G. C. Flynn, J. Biol. Chem., 2011, 286, R. J. Harris, Dev. Biol. (Basel, Switz.), 2005, 122, H. S. Gadgil, P. V. Bondarenko, G. D. Pipes, T. M. Dillon, D. Banks, J. Abel, G. R. Kleemann, and M. J. Treuheit, Anal. Biochem., 2006, 355, L. W. Dick, D. Qiu, R. B. Wong, and K. C. Cheng, Biotechnol. Bioeng., 2010, 105, L. W. Dick, D. Qiu, and K. C. Cheng, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2009, 877, A. Beck, E. Wagner-Rousset, D. Ayoub, A. Van Dorsselaer, and S. Sanglier-Cianférani, Anal. Chem., 2013, 85, S. Fekete, A. L. Gassner, S. Rudaz, J. Schappler, and D. Guillarme, TrAC, Trends Anal. Chem., 2013, 42, S. Hjerten, J. Chromatogr., 1983, 270, A. S. Cohen and B. J. Karger, J. Chromatogr., 1987, 397, J. Liu, S. Abid, and M. S. Lee, Anal. Biochem., 1995, 229, G. Hunt, K. G. Moorhouse, and A. B. Chen, J. Chromatogr. A, 1996, 744, G. Hunt and W. Nashabeh, Anal. Chem., 1999, 71, O. Salas-Solano, B. Tomlinson, S. Du, M. Parker, A. Strahan, and S. Ma, Anal. Chem., 2006, 78, D. A. Michels, M. Parker, and O. Salas-Solano, Electrophoresis, 2012, 33, M. E. Le, A. Vizel, and K. M. Hutterer, Electrophoresis, 2013, 34, M. Eggers and D. Ehrlich, Hematol. Pathol., 1995, 9, S. Yao, D. S. Anex, W. B. Caldwell, D. W. Arnold, K. B. Smith, and P. G. Schultz, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, L. Bousse, S. Mouradian, A. Minalla, H. Yee, K. Williams, and R. Dubrow, Anal. Chem., 2001, 73, J. Han and A. K. Singh, J. Chromatogr. A, 2004, 1049, H. Nagata, M. Tabuchi, K. Hirano, and Y. Baba, Electrophoresis, 2005, 26, X. Chen, K. Tang, M. Lee, and G. C. Flynn, Electrophoresis, 2008, 29, X. Chen and G. C. Flynn, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2009, 877, 3012.