Optimization and Validation of a Quantitative Capillary Electrophoresis

Aug 5, 2006 - Optimization and Validation of a Quantitative Capillary Electrophoresis Sodium Dodecyl Sulfate Method for Quality Control and Stability ...
0 downloads 4 Views 224KB Size
Download PDF
Anal. Chem. 2006, 78, 6583-6594

Optimization and Validation of a Quantitative Capillary Electrophoresis Sodium Dodecyl Sulfate Method for Quality Control and Stability Monitoring of Monoclonal Antibodies Oscar Salas-Solano,*,† Brandon Tomlinson,† Sarah Du,‡ Monica Parker,† Alex Strahan,‡ and Stacey Ma†

Departments of Late Stage Analytical Development and Quality Control Analytical Technologies, Genentech, Inc., South San Francisco, California 94080

In previous work, a capillary electrophoresis sodium dodecyl sulfate (CE-SDS) method using precolumn labeling and laser-induced fluorescence (LIF) detection was developed at Genentech Inc. as part of the control system for the quality control release of a recombinant monoclonal antibody (rMAb) (Hunt, G.; Nashabeh, W. Anal. Chem. 1999, 71, 2390-2397.). In the current work, a generic and quantitative CE-SDS assay with LIF detection of rMAbs with improved accuracy and precision is described. The implementation of an alkylating step with iodoacetamide and optimization of the incubation temperature and time, in the presence of SDS, greatly decrease any thermally induced fragmentation of nonreduced labeled rMAb samples. In addition, a quantitative study of the effects of sample buffer pH on rMAb fragmentation is also presented. Furthermore, the performance of alternative CE-SDS polymer solutions and instrumentation for quantitative analysis of rMAbs is shown in this article. The validation of this method, under the guidelines of the International Committee on Harmonization (ICH), demonstrates that the assay quantitatively determines the consistency of rMAb manufacture as it relates to size heterogeneity and product purity. In the past decade, there has been rapid progress in the development of therapeutic proteins produced using recombinant DNA technology. Among them, monoclonal antibodies have become the most rapidly growing class of biopharmaceutical products.1 Moreover, the biopharmaceutical industry largeantibody pipelines reveal the allure of recombinant monoclonal antibody (rMAbs) as therapeutics.2 With new tools in place for large-scale production of rMAbs, there is an increasing need for detailed product characterization and control of the manufacturing process. Concomitantly, an understanding of these process changes on the homogeneity, purity, and molecular stability of these recombinant proteins is also required. Therefore, analytical * Corresponding author: (e-mail) [email protected]; (fax) 650-225-3554. † Department of Late Stage Analytical Development. ‡ Department of Quality Control Analytical Technologies. (1) Wurm, F. Nat. Biotechnol. 2004, 22, 1393-1398. (2) Adams, G. P.; Weiner, L. M. Nat. Biotechnol. 2005, 23, 1147-1157. 10.1021/ac060828p CCC: $33.50 Published on Web 08/05/2006

© 2006 American Chemical Society

techniques must be developed to assess physicochemical properties such as charge-, size-, and hydrophobicity-based distribution of these complex biopharmaceutical products. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) has traditionally been used as the primary method for size-based protein separations under denaturing conditions.3 SDS binds to polypeptide chains, resulting in similar charge densities and constant mass-to-charge ratios. Hence, an electrophoretic separation of these SDS-protein complexes based on size can be achieved in a sieving medium.4 Detection of proteins separated by SDS-PAGE is generally accomplished by staining with either Coomassie Brilliant Blue5 or the more sensitive silver stain dyes.6 As the logarithm of the molecular mass of a protein is linear with electrophoretic mobility, the apparent molecular weight of a given protein can be estimated from an analysis of a series of protein standards. This SDS-PAGE methodology is also used to monitor consistency of manufacture since the size heterogeneity and purity of protein preparations can be determined. The major drawbacks of conventional SDS-PAGE are its inconvenience, the irreproducibility associated with the staining/ destaining steps used in analyte detection, the use of toxic reagents, and high intra- and intergel effective mobility variability. The emerging technique of capillary electrophoresis (CE) shows many advantages over classical SDS-PAGE including oncolumn direct UV or fluorescence detection, automation, enhanced resolution and reproducibility, and accurate quantification of proteins and molecular weight determination.7-14 Currently, replace(3) Weber, K.; Osborne, M. J. J. Biol. Chem. 1969, 244, 4406-4412. (4) Reynolds, J. A.; Tanford, C. J. J. Biol. Chem. 1970, 24, 5161-5165. (5) Chrambac, A.; Reisfeld, R. A.; Wyckoff, M.; Zaccari, J. J. Anal. Biochem. 1967, 20, 150-155. (6) Oakley, B. R.; Kirsch, D. R.; Morris, N. R. Anal. Biochem. 1980, 105. (7) Shen, Y.; Smith, R. D. Electrophoresis 2002, 23, 3106-3124. (8) Guttman, A. Electrophoresis 1996, 17, 1333-1341. (9) Schmerr, M. J.; Jenny, A.; Cutlip, R. C. J. Chromatogr., B: Biomed. Appl. 1997, 697, 223-229. (10) Guttman, A.; Nolan, J. Anal. Biochem. 1994, 221, 285-289. (11) Hu, S.; Jiang, J.; Cook, L. M.; Richards, D. P.; Horlick, L.; Wong, B.; Dovichi, N. J. Electrophoresis 2002, 23, 3136-3142. (12) Manabe, T. Electrophoresis 1999, 20, 3116-3121. (13) Dovichi, N. J.; Hu, S.; Michels, D.; Zhang, Z.-C.; Krylov, S. N. In Biotechnology, 2nd ed.; Sensen, C. W., Ed.; Wiley-VCH: Weinheim, 2001; pp 269-277. (14) Shieh, P. C. H.; Hoang, D.; Guttman, A.; Cooke, N. J. Chromatogr. 1994, 676, 219-226.

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006 6583

ableand water-soluble linear or slightly branched polymers, such as linear polyacrylamide, poly(ethylene oxide),15 poly(ethylene glycol), dextran, and pullulan are used as the sieving matrix for CE-SDS. 8,16-19 The implementation of these polymer solutions has led to the enhancement of overall precision and robustness of this CE-based methodology.20 Recently, CE-SDS has been recognized as an important tool in the biopharmaceutical industry to support analytical characterization, process development, and quality control of therapeutic rMAbs.21-28 UV absorbance is a common detector in CE-SDS; however, the sensitivity is low due to the short optical path length across the capillaries.26 Hunt and Nashabeh22 demonstrated that precolumn labeling of recombinant protein products with 5-carboxytetramethylrhodamine succinimidyl ester (5-TAMRA.SE) and laser-induced fluorescence (LIF) detection can be used for CE-SDS analyses as a replacement for silver-stained SDS-PAGE to detect low-level impurities and rMAb size variants and to determine lot-to-lot consistency. As a routine assay, qualitative analyses were employed to demonstrate that the molecular mass distribution of a bulk manufacture run was comparable to that of a well characterized reference material, similar to the practice used in silver-stained SDS-PAGE. An advantage of CE-SDS over other size-based separation methods such as high-performance size-exclusion chromatography (HPSEC) and SDS-PAGE is improved resolution of more closely related size variants of proteins.22,24 Ma27 recently demonstrated the high separation power of CE-SDS by showing baseline resolution between the heavy chain (HC) and nonglycosylated heavy chain (NGHC) species present at a low level (2%). Thus, the CE-SDS analysis of reduced rMAb samples facilitates more reliable, accurate, and reproducible quantification of the glycolysation site occupancy. Due to these features, CE-SDS is expected to be a routine technique for the quantitative analysis of a rMAb, its fragments, and nonproduct related impurities. However, Hunt and Nashabeh22 described the presence of sample preparation artifacts in the form of thermally induced fragmentation, aggregation, or both of the rMAb samples before CE-SDS analysis with LIF detection using conditions traditionally used in SDS-PAGE analysis. These issues limited the application of this assay to quantitatively determine the size distribution of therapeutic rMAbs, particularly of nonreduced samples. The induced frag(15) Verhelst, V.; Mollie, J.-P.; Campeol, F. J. Chromatogr., A 97, 770, 337-344. (16) Ganzler, K.; Greve, K.; Cohen, A. S.; Karger, B. L.; Guttman, A.; Cooke, N. C. Anal. Chem. 1992, 64, 2665-2671. (17) Wu, D.; Regnier, F. E. J. Chromatogr. 1992, 608, 349-356. (18) Nakatani, M.; Shibukawa, A.; Nakagawa, T. J. Chromatogr., A 1994, 672, 213-218. (19) Benedek, K.; Guttman, A. J. Chromatogr., A 1994, 680, 375-381. (20) Ma, S.; Nashabeh, W. Chromatographia 2001, 53, S75-S89. (21) Lee, H. G.; Chang, S.; Fritsche, E. J. Chromatogr., A 2002, 947, 143-149. (22) Hunt, G.; Nashabeh, W. Anal. Chem. 1999, 71, 2390-2397. (23) Schenerman, M. A.; Bowen, S. H. Chromatographia 2001, 53, S66-S74. (24) Krull, I. S.; Liu, X.; Dai, J.; Gendreau, C.; Li, G. J. Pharm. Biomed. Anal. 1997, 16, 377-393. (25) Tous, G. I.; Wei, Z.; Feng, J.; Bilbulian, S.; Bowen, S.; Smith, J.; Strouse, R.; McGeehan, P.; Casas-Finet, J.; Schenerman, M. A. Anal. Chem. 2005, 77, 2675-2682. (26) Patrick, J. S.; Lagu, A. L. Electrophoresis 2001, 22, 4179-4196. (27) Ma, S. In State of the Art Analytical Methods for the Characterization of Biological Products and Assessment of Comparabiity; Mire-Sluis, A. R., Ed.; Karger: Basel, 2005; pp 49-68. (28) Strong, R. A.; Liu, H.; Krull, I. S.; Cho, B.-Y.; Cohen, S. A. J. Liq. Chromatogr. Relat. Technol. 2000, 23, 1775-1803.

6584

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

mentation was attributed to trace levels of free sulfhydryl groups in the multiple disulfide-linked rMAbs.21,22,27 It was also reported that high pH conditions during heat treatment also enhanced the fragmentation of SDS-rMAb complexes.29 These artifacts significantly alter the true representation of the size heterogeneity of a protein and also increase the variability of quantitative CE-SDS methodologies. In the current work, the precolumn labeling of rMAb samples with 5-TAMRA.SE and subsequent SDS complexation conditions are further optimized to reduce these artifacts in order to quantitatively monitor the purity of rMAbs as a routine assay during drug development and quality control. As demonstrated below, the CE-SDS analysis of rhodaminelabeled rMAb samples using different separation matrixes showed comparable resolution among the various size variants observed in rMAb preparations. Moreover, sensitivity equivalent to silverstained SDS-PAGE was also obtained with the different LIF detection systems. After the optimization studies were completed and a thorough understanding of the method was obtained, validation studies were performed according to the guidelines of the International Committee on Harmonization (ICH).30 EXPERIMENTAL SECTION Instrumentation. The CE-SDS analyses were performed using a Proteomelab PA800 CE system equipped with a LIF detector (Beckman Coulter, Inc., Fullerton, CA). Briefly, the LIF detector uses a 3.5-mW argon ion laser with an excitation wavelength of 488 nm and an emission band-pass filter of 560 ( 20 nm. CE-SDS separations were performed in fused-silica capillaries (Polymicro Technologies, Phoenix, AZ) of 50-µm i.d. and 375-µm o.d. with a total length of 31.2 cm and an effective length of 21.2 cm. Other electrophoretic conditions are listed in the figure captions. The Bio-Rad BioFocus 3000 CE system with a LIF detector (Bio-Rad Laboratories, Inc. Hercules, CA) was also used. The LIF detector also uses a 3.5-mW argon ion laser with an excitation wavelength of 488 nm and an emission band-pass filter of 560 ( 20 nm. Preparation and Purification of SDS-rMAb-Labeled Conjugates. rMAb samples (0.5 mg) were buffer exchanged into 1000 µL of 0.1 M sodium bicarbonate, pH 8.3, using NAP-5 columns. Ten microliters of 5-TAMRA.SE (0.17 µg/µL) dissolved in dimethyl sulfoxide (DMSO) was then added to 190 µL of bufferexchanged rMAb solution, and the final solution was incubated for 30 min at 30 °C. For the removal of excess dye, a volume of 190 µL of the antibody-dye conjugate was loaded onto a second NAP-5 column (equilibrated with 85 mM citrate-phosphate buffer, pH 6.5), unless otherwise indicated in the Results and Discussion section. Then, 400 µL of the citrate-phosphate buffer was added as the chase volume. The labeled rMAb conjugates were collected from the NAP-5 column by adding 700 µL of the same buffer solution. Nonreduced SDS-rMAb conjugates were then prepared by mixing 150 µL of the purified labeled rMAb sample and 50 µL of a 4% (w/w) SDS solution containing 160 mM iodoacetamide (IAM). For reduced SDS-rMAb complexes, 150 µL of the purified labeled rMAb sample was mixed with 50 µL of 4% (w/w) SDS solution and 10 µL of 0.1 M DTT. The nonreduced and reduced samples were incubated at 70 °C in a water bath for 5 and 20 (29) Lee, H. G. J. Immunol. Methods 2000, 234, 71-81. (30) ICH Q2B. Guideline on validation of analytical procedures: methodology. Fed. Regist. 1997, 62, 27464-27467.

min, respectively, unless otherwise noted in Results and Discussion. Finally, a volume of 100 µL of the nonreduced or reduced sample was added to a PCR tube and placed in the sample compartment of the Proteomelab PA800 for the electrophoretic separation. Deglycosylation of rMAb Samples. Approximately 3 mg of the rMAb was treated with peptidyl N-glycosidase (PNGase F) from Prozyme (San Leandro, CA) in a solution containing 50 mM Tris, 4 mM EDTA, pH 8.0. The enzyme ratio was 30 munits/mg (1:333 E/P). The sample was digested for 18 h at 37 °C in a water bath. CE-SDS Analysis. Prior to the separation of the labeled rhuMAb samples by CE-SDS, the capillary was rinsed at 70 psi with 0.1 M NaOH, 0.1 M HCl, and deionized water for 5.0, 1.0, and 1.0 min, respectively. The CE-SDS polymer solution (or CE-SDS running buffer) was loaded into the capillary at 50 psi for 15 min from the outlet side. Samples were injected electrokinetically at 10 kV for 10 s. The CE analysis was conducted in the negative polarity mode (-15 kV, -480 V/cm). Typical current obtained was ∼32.5 µA when the capillary temperature was maintained at 40 °C. Mass Spectrometic Analysis. The LC/ESI MS analysis of nonreduced labeled rMAb was performed on a Hewlett-Packard 1090 HPLC system; it was directly interfaced into the electrospray source of a PE Sciex 3000 triple quadruple mass spectrometer (Foster City, CA). The nonreduced sample was injected onto a perfusion chromatography flow-through reversed-phase column (Poros R2 10 µm, 0.33 mm × 20 cm). The column was equilibrated at 65 °C with 0.1% formic acid at 250 µL/min. Samples were held in the column for 5 min in 0.1% formic acid and then introduced into the mass spectrometer by using a step gradient to 20%, followed by a linear gradient from 20 to 70% of a solution containing 0.08% formic acid in acetonitrile. The mass spectrometer was operated in the positive-ion mode. Data processing was done using the PE Sciex software. Reagents. Humanized rMAbs were produced in-house using transfected Chinese hamster ovary cells (Genentech, Inc., So. San Francisco, CA). The SDS gel buffer, the 0.1 N NaOH basic wash, and the 0.1 N HCl acidic wash were manufactured by Beckman Coulter, Inc. (Fullerton,CA). All chemicals used to prepare the reagents were analytical reagent grade. Tris(hydroxymethyl)aminomethane (Tris), 2-mercaptoethanol, sodium phosphate, sodium citrate, DMSO, and IAM were obtained from SigmaAldrich (St. Louis, MO). Sequencing grade SDS was purchased from Pierce (Rockford, IL). 5-TAMRA.SE was purchased from Molecular Probes (Eugene, OR). NAP-5 columns were obtained from Amersham Biosciences (Piscataway, NJ). RESULTS AND DISCUSSION Optimization of rMAb Sample Preparation Scheme for Quantitative CE-SDS Analysis with LIF Detection. rMAb Buffer Exchange. The presence of lysine residues in most proteins provides ideal reactive sites for precolumn labeling above pH 8.0.31 The use of 5-TAMRA.SE as a labeling reagent greatly increases the detection sensitivity while maintaining a genuine representation of the antibodies.22 To enhance the reliability of the labeling step, the antibody solution must be free of any amine(31) Brinkley, M. Bioconjugate Chem. 1992, 3, 2-13.

containing substances. Therefore, the antibody samples were buffer exchanged into a solution of 0.1 M sodium bicarbonate, pH 8.3, using a NAP-5 column packed with Sephadex G25. Six different 500-µL aliquots of a solution with a known concentration of rMAb were added to various NAP-5 columns previously equilibrated with 0.1 M sodium bicarbonate, pH 8.3. Then, 1-mL aliquots of the bicarbonate solution were added to elute the proteins from the columns. The final concentration of the collected rMAb samples was measured at 280 nm. Overall, the rMAb recovery was determined to be satisfactory and greater than 99.5% with a RSD of less than 0.3% (n ) 6 samples). Dye-to-Protein (D/P) Ratio. During the method development, optimization of the D/P molar ratio is important to enhanced sensitivity without the formation of rMAb aggregates or high molecular weight species (HMW). Due to the hydrophobic nature of the dye, it was reported that rhodamine-rMAb conjugates tend to aggregate more than the unlabeled antibody.22 Several rMAb samples were labeled using D/P molar ratios from 1:1 to 20:1, in increments of 5. The sample solutions were incubated in a water bath at 30 °C for 30 min. To decrease side reactions and to obtain a more consistent degree of labeling, reactions with the 5-TAMRA.SE dye should be performed at low temperatures.31 For these experiments, the excess unreactive dye removal and SDS complexation steps were performed according to the procedure described by Hunt and Nashabeh.22 Results of further optimization of these steps noted in this work are discussed below. To determine the degree of labeling of the rMAb samples at the different D/P ratios by LC/ESI MS, the antibody was first deglycosylated using PNGase F to simplify the mass spectra by minimizing the sample heterogeneity associated with the carbohydrate structures. Then, aliquots of the deglycosylated rMAb sample were labeled with 5-TAMRA.SE dye at the same D/P molar ratios (1:1-20:1) and purified as described above. Modifying a lysine group in the rMAb by the incorporation of a 5-TAMRA.SE dye molecule would increase the molecular mass of the unlabeled antibody (145 952 Da) by 413 Da. Figure 1 shows the eletrophoretic separations of nonreduced rMAb-labeled samples prepared at different D/P ratios by CE-SDS with LIF detection. The axes in this figure were offset ∼10% for clarity purposes. The major peak observed is the intact antibody. The inset presents an expanded view from 7 to 15 min and smaller peaks, attributed to various rMAb fragments with different combinations of light and heavy chains, are also seen.22 Finally, the peak migrating at the end of the electropherogram corresponds to the HMW species. Upon increasing the D/P ratio from 1:1 to 5:1, the corrected peak area (CPA), defined as the peak area divided by the migration time, of the intact antibody increased 4-fold. On the other hand, the corrected percent peak area (%CPA), defined as the CPA divided by the total corrected peak area and multiplied by 100, of the HMW species increased slightly from 0.1 to 0.2%. The aggregation level for the labeled samples prepared using a D/P ratio of 5:1 was comparable to that of the unlabeled sample as determined by native HPSEC at 0.2% (data not shown). The increase in fluorescent signal observed for the sample at a D/P ratio of 5:1 is explained by examining the reconstructive mass spectra of these two labeled antibody samples shown in Figure 2A and B. As expected, a distribution of dyeantibody conjugates with varying degrees of labeling was observed Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

6585

Figure 1. CE-SDS separations of nonreduced labeled rMAb samples prepared at different D/P molar ratios. The inset shows an expanded view of the electropherograms. Electrophoretic conditions: instrument, Bio-Rad Bio Focus 3000 capillary electrophoresis system. Detection was performed with LIF using a 3.5-mW argon ion laser, 488 nm excitation, 560 ( 20 nm emission. Effective length 19.4 cm, total length 30 cm, 50-µm-i.d., 375-µm-o.d. uncoated fused-silica capillary; both anode and cathode buffers were the Bio-Rad SDS running buffer. The samples were injected at a constant electric field of 417 V/cm for 15 s and electrophoresed at 625 V/cm (21.2 µA) at 20 °C.

for all the samples. The main species present in the rMAb-labeled sample prepared at a D/P ratio of 1:1 were the unlabeled antibody and rhodamine-rMAb conjugates with 1:3 dye molecules/protein. By increasing the D/P ratio to 5:1, the labeling of the active lysine groups in the deglycosylated rMAb was more efficient. The most prominent species present in this sample were the dye-antibody conjugates with 2-7 dye molecules/protein. Rhodamine-rMAb conjugates with up to 10 dye molecules/protein could be seen in this sample. Only a small fraction of the unlabeled rMAb species was also present in the sample. At a D/P ratio of 10:1, the CPA of the intact antibody peak was ∼1.3-fold larger than that of the sample with a D/P ratio of 5:1 (Figure 1). However, the %CPA of the aggregates for the sample at a D/P ratio of 10:1 increased 5-fold to ∼1.0%. The mass spectra for the sample, prepared at a D/P ratio of 10:1 in Figure 2C, shows that rhodamine-antibody conjugates with 3-8 dye molecules/protein were the dominant species after the labeling reaction. Moreover, the distribution of rhodamine-rMAb conjugates shifted toward the formation of labeled species with up to 12 dye molecules/protein. This could explain the significant increase in HMW species observed in this sample. It was reported that a substitution of 2-6 dye molecules/antibody was recommended in order to preserve its properties.31,32 The relative small gain in fluorescence signal upon increasing the D/P ratio from 5:1 to 10:1 is probably due to a reduction in the fluorescence efficiency as a result of quenching when tags are located in proximity to one another. This phenomenon was reported in proteins with a high degree of labeling.31 The solutions with a D/P ratio of 15:1 and 20:1 showed %CPA values for the HMW species of approximately 2 and 3.5%, respectively (Figure 1). In (32) Krull, I. S.; Strong, R.; Sosic, Z.; Cho, B.-Y.; Beale, S. C.; Wang, C.-C.; Cohen, S. J. Chromatogr., B: Biomed. Appl. 97, 699, 173-208.

6586

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

addition, the CPA of the intact antibody peak decreased by ∼10% for the solution with a D/P ratio of 15:1 when compared to the solution with a D/P ratio of 10:1. Furthermore, a decrease of ∼15% in the CPA value of the intact antibody was observed for the solution with a D/P ratio of 20:1 relative to the sample prepared at a D/P ratio of 10:1. The reconstructed mass spectrum in Figure 2D corresponding to the sample prepared at a D/P ratio of 15:1 had up to 15 dye molecules/protein. The rMAb sample labeled at a D/P ratio of 20:1 contained up to 20 dyes/antibody as seen in Figure 2E. Probably, the increase in hydrophobicity of the rMAb associated with the high degree of labeling of these samples caused the formation of significant amounts of aggregates. Additionally, the loss in fluorescence signals observed in the CE-SDS analysis of these highly labeled samples could be attributed to a more extensive quenching of the fluorescence signals. Evaluation of the data above determined that labeling of rMAb samples with 5-TAMRA.SE for CE-SDS analysis with LIF detection was optimum at a D/P ratio value of 5:1. These experimental conditions provided enhanced sensitivity comparable to silver stain SDS-PAGE (as demonstrated below) and minimized sample preparation artifacts such as aggregation. It is important to note that with exception of the dye peak, which was detected in the blank run, the CE-SDS electropherograms did not show any dyerelated artifacts, multiple peak formation from a single analyte, or loss of resolution due to the heterogeneous nature of the labeling reaction. Identical electrophoretic profiles were obtained for nonlabeled and 5-TAMRA.SE-labeled rMAb samples detected by UV and LIF, respectively (data not shown). Excess Dye Removal. It is important to remove the excess unreacted dye to effectively terminate the labeling reaction. Also, the free dye could comigrate with earlier eluting peaks in the CE-SDS profiles.22 Therefore, an experiment was performed

Figure 2. Reconstructed mass spectra of nonreduced labeled rMAb samples prepared at different D/P molar ratio: (A) 1:1; (B) 5:1; (C) 10:1; (D) 15:1; (E) 20:1. Detailed experimental conditions of the LC/ESI MS analysis are described in the Experimental Section.

to determine the elution volume of the NAP-5 columns required to efficiently collect the rhodamine-rMAb conjugates with minimum amounts of the unreacted dye. In this step, an aliquot of 190 µL of a labeled rMAb sample (D/P ratio of 5:1) was added to a NAP-5 column equilibrated with 100 mM Tris-HCl buffer (pH 9.0). A buffer aliquot of 100 µL was added to the column and then collected. This procedure was repeated to generate 15 fractions

of 100 µL of the purified labeled sample. Each of the fractions was then mixed with a SDS solution to obtain a final concentration of 1% (w/w) and incubated in a water bath at 90 °C for 3 min prior to CE-SDS analysis with LIF detection. Most of the rhodamine-rMab conjugates eluted in fractions 6-12 (data not shown). The presence of unreacted dye increased in the subsequent fractions that contained negligible amounts of the labeled Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

6587

rMAb conjugates. As a result, the procedure for optimum purification and recovery of the labeled rMAb sample included the following: (i) loading 190 µL of labeled sample to the NAP-5 column, (ii) adding 400 µL of the elution buffer (100 mM TrisHCl buffer, pH 9.0) to allow the samples to enter the gel bed, and (iii) collecting the purified labeled protein by the addition of 700 µL of the elution buffer. Incubation of 5-TAMRA Labeled rMAb Samples with SDS. Traditional sample preparation conditions to form SDS-protein complexes prior to electrophoretic analysis includes heat treatment at elevated temperatures (e.g., 90 °C).21,22,24,29,33 In the case of nonreduced rMAbs, this could lead to sample preparation artifacts in the form of thermally induced fragmentation attributed to disulfide reduction and exchange reactions.22 Moreover, it was reported that high pH (>9.0) also enhanced the fragmentation of antibodies due to the same disulfide reshuffling.21 Since high pH also facilitates the antibody complexation with SDS, a tradeoff needs to be explored. As expected, the induced fragmentation of rMAb during the sample preparation for CE-SDS analysis under nonreducing conditions depends on multiple factors. In this section, a systematic study of the effect of sample alkylation, heating temperature, incubation time, and sample buffer pH on rMAb fragmentation for CE-SDS analysis with LIF detection is discussed. (1) Alkylation of Labeled rMAb Samples. Reduction of disulfide linkages by nearby SH groups of cysteine residues and exchange reactions were proposed as an explanation for the presence of free light-chain (LC) and heavy-chain (HC) fragments in addition to other subunits of a monoclonal antibody.34,35 On the other hand, several reagents have proven to be useful for protection of thiols prior to analysis by Edman sequencing, mass spectrometry, and SDS-PAGE, including IAM and iodoacetic acid (IAA).36 As a result, an investigation of the effect of these alkylating agents on the thermally induced fragmentation of labeled nonreduced rMAb samples for CE-SDS analysis was performed. The labeled rMAb samples were reconstituted in 100 mM Tris-HCl buffer (pH 9.0) as described above. It was reported that alkylation of proteins with IAM or IAA is more efficient at pH levels higher than 8 since the cysteine reacts as a thiolate anion.36 To systematically understand the effect of alkylation on rMAb fragmentation, the pH of sample solution was kept constant at 9.0. In addition, all the nonreduced samples were incubated at 90 °C for 5 min in a water bath prior to CE-SDS analysis. The reduced samples were incubated at 90 °C for 10 min in the presence of 1 M DTT. Different amounts of IAM or IAA were dissolved in the SDS solution to final concentration values of 10, 40, 80, and 100 mM alkylating agent, which are similar to concentrations used to completely alkylate peptides during mapping procedures.36 Nonreduced SDS-rMAb conjugates were prepared by mixing the labeled rMAb samples reconstituted in 100 mM Tris-HCl buffer (pH 9.0) with the SDS solutions containing IAM or IAA at the different concentrations cited above. The final SDS concentration in all the samples was 1% (w/w). The control sample was the labeled rMAb sample in SDS solution without the alkylating agent. As shown in Figure 3A, a significant decrease of the peak areas (33) Walker, J. M. In The Protein Protocols Handbook, 2nd ed.; Human Press: Totowa, NJ, 2002; pp 61-79. (34) Li, L.; Sun, M.; Gao, Q.; Sudhir, P. Mol. Immunol. 1996, 33, 593-600. (35) Gray, W. Protein Sci. 1993, 2, 1732-1748. (36) Lundell, N.; Schreitmuller, T. Anal. Biochem. 1999, 266, 31-47.

6588

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

corresponding to the rMAb fragments was observed for the sample with 40 mM IAM in the solution compared to the control sample. The corrected percent peak area (%CPA) of the intact antibody increased from 90.0% in the nonalkylated sample to 95.0% in the sample containing 40 mM IAM. Similar results were obtained when IAA was used as the alkylating agent. However, IAM was the reagent of choice since solutions of IAA were less stable as indicated by the appearance of a yellow coloration after a few minutes. It was also determined that the optimum concentration of IAM in the sample was 40 mM. At concentrations lower than 40 mM IAM, some induced fragmentation was still observed (data not shown). On the other hand, the presence of IAM in the sample solutions at concentrations higher than 40 mM did not further reduce fragmentation. The suppression of rMAb fragmentation upon alkylation implicates free SH groups present in rMAbs as the mediators of the disulfide exchange reactions for the formation of antibody fragments. It is important to mention that the aggregation level of the alkylated, labeled rMAb samples was similar to that of the nonalkylated control sample (0.2%). Also, the addition of the alkylating agent to reduced labeled rMAb samples was not required. As shown in Figure 3B, upon the addition of 1 M DTT, the rMAb is reduced to LC, NGHC, HC, and other small antibody fragments. The CE-SDS profiles of the reduced samples prepared with 40 mM IAM or IAA were comparable to that of the control sample that did not contain the alkylating agent. Similar results were obtained for the reduced labeled rMAb samples containing other concentrations of IAM or IAA. (2) Effect of Incubation Temperature and Time. The effect of temperature (45, 60, 70, and 90 °C) and heating time (5, 10, 15, and 20 min) on rMAb fragmentation in the absence or presence of the alkylating agent (40 mM IAM) before CE-SDS analysis of nonreduced samples was studied. The pH level of these sample solutions for SDS-rMAb complexation was pH 9.0. Figure 4A shows the %CPA of the intact antibody as a function of heating time for rMAb-labeled samples, without IAM, incubated at different temperatures using a water bath. The extent of rMAb fragmentation at the various incubation times was more pronounced upon increasing the temperature. For example, the %CPA of the intact antibody significantly decreased from 97 to 90% by increasing the water bath temperature from 45 to 90 °C for those labeled rMAb samples incubated for 5 min. It is important to note that a traditional sample preparation scheme for SDS-PAGE and CE-SDS includes temperatures higher than 90 °C.21,22,24,29,33 For all temperatures studied, the %CPA of the intact antibody decreased with longer heating times. On the other hand, it can be observed in Figure 4B that, with the addition of IAM to the labeled rMAb samples, the thermally induced fragmentation of the intact antibody was significantly suppressed. The %CPA of the intact antibody for labeled rMAb samples, heated in the presence of IAM, at 45, 60, and 70 °C remained at 97% even after a 20-min incubation at those temperatures. The curve at 90 °C showed an increase in fragmentation over time. However, the rate of %CPA change for these alkylated samples was significantly lower at 4% after 20 min than that observed for the samples without IAM incubated under otherwise the same experimental conditions. Finally, the proposed alkylation of free thiols present in the samples not only decreased rMAb fragmentation but also signifi-

Figure 3. Expanded view of CE-SDS separations of (A) nonreduced and (B) reduced labeled rMAb samples in the presence of different alkylating agents. Electrophoretic conditions: instrument, Bio-Rad Bio Focus 3000 capillary electrophoresis system. Detection was performed with LIF using a 3.5-mW argon ion laser, 488 nm excitation, 560 ( 20 nm emission. Effective length 19.4 cm, total length 30 cm, 50-µm-i.d., 375-µm-o.d. uncoated fused-silica capillary; both anode and cathode buffers were the Bio-Rad SDS running buffer. The samples were injected at a constant electric field of 417 V/cm for 15 s and electrophoresed at 625 V/cm (21.2 µA) at 20 °C.

cantly reduced the standard deviation of the %CPA values of the intact antibody at the various incubation temperatures as shown in Figure 4B. Taken together, these studies demonstrate that alkylation of the rMAb samples during the heating step for protein-SDS complexation significantly improved the robustness and accuracy of the assay. An important aspect of quantitative analysis by CE-SDS is to obtain a homogeneous rMAb peak for accurate and reproducible integration. All the nonreduced, labeled rMAb samples heated at 45 °C showed a broad peak for the intact antibody. Increasing

the heating time of samples up to 20 min did not eliminate this phenomenon (data not shown). Upon increasing the incubation temperature to 60 °C, a sharper peak for the intact antibody was obtained; however, a shoulder at the tailing edge of the peak remained in all the samples. On the other hand, incubation of the nonreduced labeled rMAb samples at 70 °C resulted in a symmetric and uniform peak for the intact antibody-SDS complexes. Uniform peak shapes for the intact antibody were also observed by heating the nonreduced samples at 90 °C; however, excessive heating proved detrimental and caused fragmentation Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

6589

Figure 4. Corrected percent peak area of intact antibody vs heating time at several incubation temperatures: (1) 45, (2) 60, (b) 70, and (9) 90 °C. (A) no IAM; (B) with IAM. Error bars are shown at the 95% confidence interval (n ) 3). Other experimental conditions as in Figure 3.

Figure 5. Corrected percent peak area of intact antibody as a function of the buffer pH of the labeled rMAb sample solution, (2) with or (b) without IAM, prior to heat treatment for CE-SDS analysis with LIF detection. Error bars are shown at the 95% confidence interval (n ) 3). Other experimental conditions as in Figure 3.

of the intact antibody (as shown in Figure 4). Based on the data presented above, the optimum experimental conditions for reproducible quantification of nonreduced labeled rMAb samples by CE-SDS analysis with LIF detection required incubation at 70 °C for 5 min in the presence of 40 mM IAM. (3) Effect of CE-SDS Sample Buffer pH. The %CPA of the intact antibody peak of nonreduced labeled rMAb samples prepared by using the optimized conditions described above was ∼97.0%. However, when the rMAb samples were analyzed under native conditions using HPSEC, the %CPA for the intact antibody was 99.5%, and the fragments were present at a ∼0.3% level, which is lower than the value observed in the CE-SDS analysis (2.8%). It is important to note that the absence of process-related 6590 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

impurities in these rMAb samples was previously established by SDS-PAGE with silver and SyproRuby staining and peptide mass mapping (data not shown). It was therefore relevant to investigate whether the fragments were related to CE-SDS assay artifacts due to the high pH of the sample solution (pH 9.0) or were indeed present in the sample. It is possible that CE-SDS offers more accurate quantification of the rMAb size variants as a result of the better resolution between the fragments and the intact antibody compared to that obtained by HPSEC. It was reported that intramolecular thiol/disulfide rearrangements were faster (on the millisecond time scale) than alkylation reactions and that using a high concentration of alkylating agent did not ensure adequate suppression of the thiol/disulfide

Figure 6. CE-SDS separations of (A) nonreduced and (B) reduced preparations of a labeled rMAb sample. Electrophoretic conditions: ProteomeLab PA800 instrument with LIF detection, effective length 21.2 cm, total length 31.2 cm, 50-µm-i.d., 375-µm-o.d. uncoated fused-silica capillary; both anode and cathode buffers were the Beckman CE-SDS polymer solution. The samples were injected at a constant electric field of 160 V/cm for 20 s and electrophoresed at 480 V/cm (32.5 µA) at 40 °C. Other conditions are described in the Experimental Section.

exchange.36 Another common method to avoid thiol/disulfide exchange reactions besides alkylation of free thiols is to keep the sample solution at acidic conditions.37 To determine the effect of sample buffer pH on rMAb fragmentation during CE-SDS analysis of nonreduced samples, several labeled rMAb aliquots of the same sample were purified and reconstituted in the following buffer solutions using NAP-5 columns: (i) 60 mM citrate-phosphate buffer, pH 3.0, (ii) 70 mM citrate-phosphate buffer, pH 4.2, (iii) 75 mM citrate-phosphate buffer, pH 5.0, (iv) 85 mM citratephosphate buffer, pH 6.5, (v) 100 mM Tris-HCl buffer, pH 8.0, and (vi) 100 mM Tris-HCl buffer, pH 9.0. The rMAb samples were

also incubated with and without 40 mM IAM, in the presence of 1% (w/w) SDS, at 70 °C for 5 min prior to the electrophoretic analysis. Figure 5 shows the %CPA of the intact antibody as a function of sample buffer pH in the presence or absence of the alkylating agent. An aggregation level of ∼0.2% was observed for all the labeled rMAb sample solutions. For the samples incubated in the presence of 40 mM IAM, the %CPA value of the intact antibody was ∼97% at pH 9.0. Upon decreasing the pH level of the sample solution to 8.0, the %CPA of the intact antibody (37) Gilbert, H. F. Methods Enzymol. 1995, 251, 8-28.

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

6591

increased to 97.5%. A value for the %CPA of the intact antibody of 98.2% was obtained by reducing the pH of the sample solution to 6.5. A further decrease of the pH levels to 5.0, 4.2, and 3.0, respectively, did not change the %CPA value of the intact antibody. However, the electropherograms of these labeled rMAb samples reconstituted in buffer solutions at pH levels lower than 6.5 showed broader peaks for the intact antibody and its size variants (data not shown). This is probably due to lower stability of SDSprotein complexes at acidic conditions, consistent with earlier reports that protein samples at acidic pH showed broader bands during SDS-PAGE analysis.38 Figure 5 also shows that nonreduced labeled rMAb samples, with or without IAM, had similar %CPA for the intact antibody when sample preparation buffers at pH levels of 3.0 and 4.2 were used. This suggested that thiol/disulfide exchange reactions are largely inhibited at pH