Charge Heterogeneity of Monoclonal Antibodies ... - ACS Publications

May 23, 2012 - ... Hubert Hertenberger , Andreas Wolfert , Christian Spick , Wilma Lau , Georg Drabner , Ulrike Reiff , Hans Koll , Apollon Papadimitr...
1 downloads 0 Views 365KB Size
Article pubs.acs.org/ac

Charge Heterogeneity of Monoclonal Antibodies by Multiplexed Imaged Capillary Isoelectric Focusing Immunoassay with Chemiluminescence Detection David A. Michels,*,† Andrea W. Tu,‡ Will McElroy,† David Voehringer,‡ and Oscar Salas-Solano§ †

Department of Protein Analytical Chemistry, Genentech, a Member of the Roche Group, 1 DNA Way, South San Francisco, California 94080, United States ‡ ProteinSimple, 3040 Oakmead Village Drive, Santa Clara, California 95051, United States § Department of Analytical Biochemistry, Seattle Genetics, Inc., 21823 30th Drive SE, Bothell, Washington 98021, United States S Supporting Information *

ABSTRACT: Characterization of charge heterogeneity of recombinant monoclonal antibodies (mAbs) requires high throughput analytical methods to support clone selection and formulation screens. We applied the NanoPro technology to rapidly measure relative charge distribution of mAbs in early stage process development. The NanoPro is a multiplexed capillary-based isoelectric immunoassay with whole-column imaging detection. This assay offers specificity, speed and sensitivity advantages over conventional capillary isoelectric focusing (CIEF) platforms. After CIEF, charge variants are photochemically immobilized to the wall of a short coated capillary. Once immobilized, mAbs are probed using a secondary antiIgG conjugated with horseradish peroxidase. After flushing away excess reagents, secondary antibodies bound to their targets are then detected by chemiluminescence upon incubation with peroxidase reactive substrates. Charge heterogeneity as determined by chemiluminescence was similar to that measured by conventional CIEF technology with absorbance detection for purified mAbs and contaminated mAbs derived directly from host cellular extract. Upon method optimization, the automated CIEF immunoassay was applied to several mAbs of varying isoelectric points, demonstrating the suitability of NanoPro as a rugged high-throughput product characterization tool. Furthermore, qualification of detection sensitivity, precision, and dynamic range are reported with discussion of its advantages as an alternative approach to rapidly characterize charge variants during process development of mAbs. Selection of the final clone is based upon product quality attributes such as aggregation, titer, and charge heterogeneity; whereas the selection of the final formulation depends strongly on the buffer conditions most robust for product stability, which often requires subjecting the representative IgG to various combinations of buffers, pH and salt concentration. These multivariate parameter studies routinely generate hundreds of samples needing immediate analysis, which presents throughput challenges for current analytical technologies. Charge heterogeneity of proteins is routinely quantified by capillary isoelectric focusing (CIEF) with ultraviolet (UV) detection at a fixed distance from the inlet of the capillary or imaging across the whole-column of the separation capillary.2,3 Within the last five years, imaged capillary isoelectric focusing (iCIEF) has become an essential tool in the pharmaceutical industry to monitor protein charge heterogeneity to assess

D

evelopment and manufacture of recombinant monoclonal antibodies (mAbs) comprises multiple process steps each capable of influencing purity of the therapeutic protein drug product. Analytical tools that characterize product variants by either their size or charge are crucial for defining purity, detecting chemical and post-translational modifications, and determining stability during long-term storage. Regulatory agencies consider charge heterogeneity to be particularly critical due to its potential impact on product pharmacokinetics and biological activity observed among acidic and basic isoforms, and provides the sponsor with increased understanding of product stability.1 It is essential for the biopharmaceutical industry to develop appropriate analytical tools for monitoring purity characteristics, such as size and charge heterogeneity. With the significant increase in the number of therapeutic targets of the biopharmaceutical industry pipeline, the need for highthroughput analytical systems has gained considerable interest from process development groups including Cell Culture and Formulation who depend on analytics during clone cell selection or product stability optimization, respectively. © 2012 American Chemical Society

Received: March 30, 2012 Accepted: May 23, 2012 Published: May 23, 2012 5380

dx.doi.org/10.1021/ac3008847 | Anal. Chem. 2012, 84, 5380−5386

Analytical Chemistry

Article

(CHO) cells and purified at Genentech (South San Francisco, CA). Proteins were thawed from −70 °C prior to use. Protein lysate from harvested cell culture fluid (HCCF) was prepared as previously described.11 Lyophilized goat antihuman IgGHRP from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA) was reconstituted in Milli-Q water (Billerica, MA), aliquoted, and stored at −80 °C. Carboxypepsidase B from human pancreas was purchased from Roche. Urea, arginine, and Pharmalytes pH 5−8 and pH 8−10.5 were obtained from Sigma-Aldrich (St. Louis, MO). Solutions of 1 M sodium hydroxide (NaOH) and 0.1 M hydrochloric acid (HCl) reagents were purchased from J.T. Baker (Phillipsburg, NJ). Methylcellulose, anolyte, and catholyte were from ProteinSimple (Toronto, ON). Capillaries, sample plates, antibody dilution buffer, secondary antibody, wash buffer, fluorescent pI standards, and luminol/peroxide were acquired from ProteinSimple (Santa Clara, CA). Carboxypeptidase B (CpB) Reaction. C-terminal lysine residues of CHO-derived mAb products were removed through CpB enzymatic digestion.12 Protein solutions were diluted to 1.25 mg/mL with water and mixed with CpB enzyme at a ratio of 100:1. The mixture was incubated at 37 °C for 20 min. Imaged Capillary Isoelectric Focusing (CIEF). Imaged CIEF was performed on an iCE280 Analyzer (ProteinSimple, Toronto, ON) equipped with a fabricated cartridge consisting of a 5-cm ×100-μm i.d. fluorocarbon-coated capillary connected to electrolytes through hollow dialysis membranes.13,14 The anolyte was 80 mM phosphoric acid and the catholyte was 100 mM NaOH each prepared with 0.1% methylcellulose. An IgG sample was prepared by diluting in water containing CpB enzyme at a ratio of 100:1 (w/w) and incubated at 37 °C for 20 min. An ampholyte mixture containing 3.1% (v/v) Pharmalytes, 2.5 M urea, 0.1% (v/v) methylcellulose, and pI 5.5 and 9.77 markers was combined with CpB-treated antibody to obtain a final concentration of 0.25 mg/mL. Final solutions were mixed and centrifuged at 11200-g for 2 min. Samples were introduced to the capillary through a Prince autoinjector and focused for 1 and 10 min at 1.5 kV and 3.0 kV, respectively. Electropherogram were imaged with the optical absorption detector operated at 280 nm. CIEF Immunoassay. Automated immunoassay cycles were performed on a NanoPro system (ProteinSimple) using twelve single-use, disposable 5-cm ×100-μm i.d. silica glass capillaries manufactured with a clear external Teflon coating. The inner wall of each capillary is coated with a proprietary photoreactive layer used for covalent and nonspecific protein binding upon exposure to UV light. A 384-well sample plate was manually prepared containing solutions of sample mixtures, secondary antibody, and luminol peroxide. Sample mixtures containing 0.1−0.5 μg/mL protein, 0.3% methylcellulose, 2 M urea, 2.5 mM arginine, fluorescent peptide pI markers (pI 4.9, 7.3 and 9.7), and 2.5% (v/v) ampholyte (30% pH 5−8 and 70% pH 8−10.5 Pharmalytes) were plated in one row. Antihuman IgG-HRP secondary antibody was diluted 1:300 into antibody dilution buffer and loaded into a second row while luminol/peroxide was mixed at a 1:1 ratio and loaded into a third row. The fully automated CIEF immunoassay was initiated after the sample plate was transferred to the autosampler chilled at 4 °C with a peltier cooler device for the duration of the run. The NanoPro system maneuvered capillaries between five different trays (one separation, two incubation, one reagent, and one sample tray) using a gripper. Capillaries were filled

downstream process consistency, product quality, stability and purity of therapeutic proteins.4−6 In quality control, validated iCIEF assays are primarily used in monitoring lot-to-lot consistency and stability testing during the manufacturing stage, which usually requires testing a few samples per batch. In contrast, cell culture and formulation studies have scaled down process development as a means to accelerate optimization, minimize costs and reduce sample handling burdens, subsequently producing hundreds of in-process samples with limited amounts of protein. Furthermore, characterization studies often require enrichment of variants collected from a variety of fractionizing devices. Such small-scale screening and characterization studies are driving the demand for higher throughput or better sensitivity. The NanoPro technology is a sensitive 12-capillary array immunoassay, which has been optimized for cell signaling analysis by protein phosphorylation using whole-capillary chemiluminescence detection.7,8 In this approach, charge protein isoforms are first separated by CIEF in a 5-cm capillary derivatized with UV-activated benzophenone.9,10 Light emitting diodes enable real-time monitoring of isoelectric focusing through excitation of fluorescent peptide internal pI standards mixed with the ampholyte-protein solution. Photochemical activation is used to capture separated protein to the inner lining of the capillary wall. After rinsing excess reagents from the capillary the protein immobilized to the wall is probed with a primary antibody as a means of specific detection of target proteins. The protein−antibody complexes immobilized to the capillary are further complexed with secondary HRP-antibody and subsequently treated with chemiluminescence reagents, which produce light within the entire capillary that can be scanned onto a CCD camera based imager. We applied this technology for the first time to measure charge heterogeneity of mAb products (Figure 1). In our approach, addition of the

Figure 1. Schematic representation of luminol chemiluminescence reaction used in the CIEF immunoassay. After separation by isoelectric focusing, mAb protein charge variants immobilize to the inner capillary wall upon a 60−80 s exposure to UV light. A secondary antibody-HRP conjugate binds to captured mAb products. Excess peroxide reacts with the HRP catalyst to generate light, which is then detected by a CCD camera imaging the full length of multiple capillaries simultaneously.

primary antibody was eliminated because direct binding of the mAb to the secondary was feasible. After optimization and qualification, we describe applications of this novel highthroughput CIEF immunoassay for practical use in process development of therapeutic monoclonal antibody products.



MATERIALS AND METHODS Materials and Reagents. Recombinant IgG1 monoclonal antibodies were manufactured in Chinese hamster ovary 5381

dx.doi.org/10.1021/ac3008847 | Anal. Chem. 2012, 84, 5380−5386

Analytical Chemistry

Article

from the separation of a pI ladder using the iCE280 Analyzer based on 280 nm absorbance detection. Resolution was optimized using ampholytic solutions of either pH 3−10, 5− 8, or 8−10.5 ampholytes mixed in various ratios. Characterization of the various ampholyte solutions through pI calibration led to a broad-range composition consisting of 30% pH 5−8 and 70% pH 8−10.5 ampholytes and having a linear pH gradient across the range of pH 5.5−9.8 based on regression analysis of apparent pI and pixel position (Supporting Information). With the addition of 2 M urea to the ampholytes, the broadrange iCIEF assay was capable of separating multiple mAb products with apparent pI values in the range of 6.0−9.5 including those prone to precipitate. Figure 2A shows the

with sample solutions by application of a vacuum for 8 s before being transferred to the separation tray where isoelectric focusing was initiated upon application of 1.5 kV for 1 min followed by 3 kV for 7 min. After the isoelectric focusing step, capillaries were exposed to UV light for 70−80 s to enable activation of the proprietary coating, thus immobilizing proteins to the capillary wall. Fluorescent pI markers excited by 532 nm green LEDs were used to map the pH gradient and provide a metric with which to align charge isoforms of the proteins of interest based on its apparent pI value. Following UV capture, excess sample solution was washed away from each capillary twice with wash buffer before incubating in antihuman IgGHRP solution for 20 min. Capillaries were then washed two more times before exposure to luminol/peroxide. The HRP conjugated to the secondary antibody catalyzed the oxidation of luminol resulting in a luminescent light emission. Capillaries were transferred to the separation tray where luminescence (120−480s) and fluorescence images were collected using a CCD camera. Peak integration and pI marker calibration (for peak alignment) were performed using Compass, version 1.3.7, software.



RESULTS AND DISCUSSION Selection of the Secondary HRP-Conjugated Antibody. The CIEF immunoassay works by detecting immobilized pI-separated protein indirectly through a secondary antibody (Ab) conjugated with horseradish peroxidase (HRP) similar to those commonly used in immunochemistry, Western Blotting, and ELISA.15 The anti-Ab-HRP conjugate is assumed to have high affinity to the Fc region of an IgG protein. In practice however, indirect detection of protein could potentially create bias with different charge isoforms depending on their affinity for the secondary antibody. For this reason, several secondary Ab candidates were screened and reported in the Supporting Information section. The charge variant profiles were compared for a mAb protein that was detected using different secondary Abs harvested from either mouse, rabbit or goat. Remarkably each secondary Ab resulted in similar binding and detection performance. The mechanism of protein-capillary immobilization was assumed to be nonspecific and would not disrupt secondary Ab binding due to limited Fc exposure after immobilization. Our observations reinforce this assumption considering the abundant amount of luminescence generated from the binding between secondary Ab and immobilized protein. Furthermore, the absence of the potential bias was demonstrated by comparing quantification of the CIEF immunoassay to that obtained by iCIEF with UV detection. As discussed below, the percent distribution of the charged isoforms of various mAb proteins between UV and chemiluminescence were similar. These results demonstrate the lack of any detectable binding bias of the secondary Ab to various immobilized charged variants. Although only minor differences were noted between the various secondary Ab, the goat Ab was selected for method development because of its enhanced ability in detecting minor charge isoforms. Broad-Range pH Gradient for Multiproduct iCIEF Separations. The use of platform or multiproduct chargebased assays is essential when supporting process development of mAb products because of the resources saved without having to reoptimize method parameters for each product. Therefore, development of a multiproduct iCIEF method capable of separating products of varying pI was highly desirable. Optimization of a broad-range pH gradient was determined

Figure 2. Detection of separated IgG charge variants by imaged CIEF with (A) absorbance and (B) chemiluminescence. The reagent blank shows that there are no components in the sample formulation that interfere with detection. Expanded views illustrate the baseline noise difference between detectors (insets). Signal-to-noise (S/N) measured for the basic peak was 2.5-fold higher for the immunoassay although the protein concentration is 1000-fold less. Experimental conditions are described in Materials and Methods.

separation of charge variants after CpB digestion for one of several mAb products evaluated. CpB digestion is necessary to simplify product charge distribution because of the inherited variability during batch-to-batch bioprocessing production. Performance was conducted by measuring variation of the relative peak distribution of total acidic variants, main peak and total basic variants across multiple sample preparations and 5382

dx.doi.org/10.1021/ac3008847 | Anal. Chem. 2012, 84, 5380−5386

Analytical Chemistry

Article

signal unless controlled through offline photobleaching, greatly challenging its robustness in practical use.19−21 In contrast, the CIEF immunoassay removes excess ampholytes from the capillary before luminescent detection essentially eliminating background noise. Because of the low background noise, improved baseline performance resulted in higher signal-tonoise (S/N) ratios for the minor peaks, as indicated by the basic variant peak in the inset of Figure 2B, favoring the CIEF immunoassay for use in characterization of samples with limited protein quantity. Reproducibility. To the best of our knowledge, this is the first CIEF immunoassay application for the determination of charge heterogeneity of mAb protein. Due to limited experience of the photoactivated coated capillaries for this application, it was important to assess interday precision of the CIEF immunoassay using the optimized broad-range ampholytes. Reproducibility was evaluated across three days with twelve sample preparations analyzed per cycle per day (Table 1). Representative profiles for a mAb product obtained for one cycle (Figure 3) were similar when tested on two additional days. The overlay demonstrates that capillary-to-capillary performance was reproducible in apparent pI calibration (RSD < 0.06% for the main peak) and relative charge isoform distribution. As shown in Table 1, the RSD value of the relative peak distribution for each cycle was