Automating a Critical Step in Glycoprotein Analysis for Vaccine Manufacture

Cavan/stock.adobe.com

Glycosylation is a critical quality attribute (CQA) for therapeutic proteins, influencing their stability, bioactivity, and immunogenicity. Glycoprotein characterization and glycosylation profiling are therefore fundamental in the development and production of biopharmaceutical molecules associated with vaccine research and manufacturing. In this rapidly growing sector, fast, reliable, and reproducible analytical methods are an absolute necessity.

One of the most established and effective techniques for separating and quantifying complex carbohydrates is high-performance anion-exchange chromatography with pulsed amperometric detection (HPAE-PAD). Although widely used, HPAE-PAD has a drawback in that traditionally it involves manual preparation of the eluent. Replacing this step with an automated approach is improving both method precision and productivity.

Principles and benefits of HPAE-PAD

HPAE-PAD is applicable to the analysis of a wide range of carbohydrates, delivering high-resolution separations and sensitive direct detection of low-picomole quantities. HPAE chromatography of neutral carbohydrates is possible because they are weakly acidic and ionize at high pH, enabling their separation as anions. Specifically developed polymeric anion-exchange columns are needed because classical silica-based high-performance liquid chromatography (HPLC) columns degrade rapidly at high pH.

Pulsed amperometry, used for detection, measures the electrical current generated by the oxidation of carbohydrates at the surface of a gold working electrode. The current is proportional to the carbohydrate concentration. If just a single potential is applied to the electrode, oxidation products gradually foul its surface, resulting in loss of signal. To prevent this, the electrode surface is cleaned by a series of potentials applied for fixed periods after the detection potential, which is referred to as a waveform. Repeated application of a waveform is the basis of pulsed amperometry.

HPAE-PAD is selective and specific for carbohydrates because pulsed amperometry detects only those compounds with functional groups oxidizable at the detection voltage applied. It enables determination of carbohydrates using small sample volumes and without laborious sample derivatization.

Addressing the key workflow challenge

Manual preparation of the high-pH eluents needed for HPAE-PAD is both time-consuming and complex. Because most HPAE-PAD problems arise from improper eluent preparation, this preparation is a critical step in ensuring analytical success (1,2).

Most HPAE-PAD eluents are composed of water, sodium hydroxide (NaOH) or potassium hydroxide (KOH), and sodium acetate. Each component is prone to issues of quality and variability, and manual preparation introduces further inconsistencies. Improperly prepared eluents can lead to serious problems in the workflow, including a high detection background, high noise, loss of analyte retention, and potential contamination of the system. Results may be poorly reproducible and make method transfer challenging.

Automated electrolytic generation of the eluent within the chromatography system offers an alternative to manual preparation. Here, an eluent generator equipped with a KOH cartridge is connected in series with one that has a methanesulfonic acid (MSA) cartridge to produce eluents containing KOH, KOH and potassium methanesulfonate (KMSA), or methanesulfonate (MSA). This dual eluent generation cartridge (EGC) approach produces KOH/KMSA eluents capable of executing many applications that use NaOH/sodium acetate. Cartridges are installed within the chromatography system and require only the addition of deionized water. System software manages the eluent concentrations and gradients, while chemicals are carefully controlled with no manual handling and, therefore, the avoidance of human error.

The following case studies illustrate the benefits of automated eluent preparation and the use of reduced volume chromatography columns in HPAE-PAD workflows for typical glycoprotein applications: sialic acid determination and profiling of N-linked oligosaccharides.

Evaluation of automated eluent preparation

Sialic acids are derivatives of the nine-carbon sugar neuraminic acid. Two of the most important are N-acetylneuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc), both of which occur in many biomolecules. In recombinant therapeutic glycoproteins, sialylation can maintain certain properties of the native protein, such as circulatory half-life, biological activity, and solubility. Because of this, increased sialylation is often engineered into recombinant molecules.

Humans generally do not produce Neu5Gc, so its presence in a therapeutic glycoprotein is potentially immunogenic. It is therefore vital to understand the sialic acid content of a glycoprotein when examining its therapeutic efficacy and safety. HPAE-PAD is widely used in this application, with analysis performed following acid hydrolysis or enzyme digestion of the glycoprotein.

In this study, modification of an established HPAE-PAD workflow for sialic acid determination replaced conventional manual eluent preparation with automated eluent generation. It also included a smaller chromatography column with a lower flow rate than that described in the conventional workflow, reducing the amount of eluent required. The modified workflow was used to determine the sialic acid content of three glycoprotein samples, evaluating the key performance parameters of separation, linearity, limits of detection (LODs), accuracy, and precision.

Setup and results. The analysis setup included a high-performance ion chromatography system (Dionex ICS-6000 HPIC, Thermo Scientific) equipped with an automated eluent generation module (Dionex ICS-6000 EG Eluent Generator, Thermo Scientific) and an electrochemical detector for pulsed amperometric detection (Dionex ICS-6000 DC Detector/Chromatography module, Thermo Scientific). Figure 1 illustrates the workflow for HPAE-PAD in dual EGC mode. The inclusion of a 1-mm chromatography column (Dionex CarboPac PA20-1 mm, Thermo Scientific) designed to work with the dual EGC mode reduces eluent consumption compared with the more conventional 3-mm column. Samples analyzed were bovine apo-transferrin (b. apo-transferrin), calf fetuin (fetuin), and human –acid glycoprotein (h. AGP), all of which underwent triplicate acetic acid hydrolysis before measurement. The full experimental set up and chromatographic conditions are described in a technical application note (3).

All figures are courtesy of the author. Figure 1. Example illustration of a high-performance ion chromatography system (HPIC) system in dual eluent generation cartridge (EGC) mode.

Changing from manually prepared NaOH/sodium acetate eluent to KOH/KMSA delivered in dual EGC mode requires new eluent conditions to achieve similar separations. Table I shows the KOH/KMSA gradient established to enable Neu5Gc separation, column washing, and column re-equilibration. This gradient gave clear separation and easy quantification of Neu5Ac and Neu5Gc peaks using the new 1-mm column as well as good separation of the Neu5Ac peak from the void (important when analyzing acid-hydrolyzed compounds that might contain other poorly retained compounds). Neu5Ac retention time matched that of the conventional method, while Neu5Gc eluted around seven minutes after; both were stable over 15 days.

Table I. Potassium hydroxide (KOH)/potassium methanesulfonate (KMSA) eluent gradient.

Calibrations for Neu5Ac and Neu5Gc were linear over the ranges studied. LODs were 20.9 nM (8.36 fmol) and 10.3 nM (4.13 fmol) for Neu5Ac and Neu5Gc, respectively, with limits of quantitation (LOQs) being 69.7 nM (27.9 fmol) and 34.4 nM (13.8 fmol).

Figure 2 shows separation of sialic acids in the glycoprotein samples tested. Neu5Ac is well separated from early eluting components and is present in all three glycoproteins. As expected, Neu5Gc is absent from the human glyco-protein (S3). The sialic acid levels in all the samples are consistent with determinations made using conventional (manual eluent preparation) HPAE-PAD.

Figure 2. Sialic acid determination of three glycoprotein acid hydrolysates. Neu5Ac is N-acetylneuraminic acid. Neu5Gc is N-glycolylneuraminic acid. ND is no data.

Recovery studies to evaluate method accuracy involved spiking sialic acids into each sample hydrolysate and a reagent blank. Recovery ranged from 95.1% to 105%, indicating that accurate sialic acid measurements were being made without column or detector overload.

Inter-day precision, determined by triplicate injection of a calibration standard on three separate days, showed variation in peak area precision of just 0.83% to 1.04% and in retention time precision of < 0.2% for all target sialic acids.

Outcomes. This study showed that the conventional HPAE-PAD method for determining sialic acids in glycoproteins can be modified to include automated eluent generation together with a 1-mm chromatography column in place of manual eluent preparation and a 3-mm column. The modified method gives excellent separation, linearity, sensitivity, accuracy, and reproducibility, and delivers similar resolution of sialic acids to the conventional method. It has the benefit of simpler operation and improved retention time precision.

Testing dual EGC

Oligosaccharides are identified based on comparison to known standards in combination with exoglycosidase digestions. Rohrer et al. (4) reported on the HPAE-PAD method giving superior resolution of immunoglobulin G (IgG) oligosaccharide separation using manually prepared NaOH/sodium acetate eluent gradients. The work summarized here uses automated eluent generation and a 1-mm column for the same analysis.

Setup and results. The general equipment setup and workflow are as shown in Figure 1 for the sialic acid study, except that a different chromatography column (Dionex CarboPac PA200 1mm, Thermo Fisher Scientific) was used. The full experimental set up and chromatographic conditions are described in detail in a technical application note (5).

Testing the separation of IgG oligosaccharides in the dual EGC setup with a 1-mm column used a solution of the most abundant uncharged (neutral) N-linked oligosaccharides and the major sialic acid-containing (charged) oligosaccharides released from human serum IgG. The solution was prepared by treating human serum IgG with rapid PNGase with and without denaturation. The glycan profile without denaturation shown in Figure 3 matches separations reported previously (2) and indicates good resolution of uncharged IgG oligosaccharides (glycans), comparable with traditional HPAE-PAD separations.

Figure 3. Separation of human serum immunoglobulin G (IgG) N-linked oligosaccharides using dual eluent generation cartridge. (A) PNGase F digest of human serum IgG (B) N-acetylneuraminic acid (Neu5Ac) standard (9.68 µm) (C) A1F standard (D) G1F standard.

The identities of the neutral and charged glycans in Figure 3 were confirmed by the appearance or disappearance of peaks following appropriate enzyme treatment. Neu5Ac presence is confirmed with neuraminidase treatment. Treatment of the glycans released from human serum IgG with neuraminidase resulted in the removal of all the major peaks in the charged glycan region of the chromatogram with a concomitant increase in neutral glycans, suggesting all the charged glycans are sialylated.

Sequential digestion with specific exoglycosidases enables further analysis of the N-linked oligosaccharides released from a glycoprotein. HPAE-PAD analysis is highly selective for the linkage positions, and component monosaccharides are useful in resolving mixtures of complex structures. Additionally, retention times are dependent on the presence or absence of peripheral monosaccharides.

Outcomes. HPAE-PAD with automated eluent generation for the separation of N-linked oligosaccharides released from glycoproteins is comparable to using manual eluent generation, and the system’s consistent performance supports improved reproducibility. Furthermore, where eluent concentration and gradient conditions require frequent optimization for different sample types, automated eluent generation can significantly reduce method development time.

Driving productivity through method improvement

Fast, reliable glycoprotein analysis is essential to meet immunity screening requirements in vaccine, gene, and cell therapy production. HPAE-PAD is the technique used most widely in the biopharmaceutical industry to separate and quantify glycans present on glycoproteins; however, the need for complex manual preparation of the high pH eluents involved leads to inconsistencies and hampers productivity. Incorrect eluent preparation is recognized as the main source of HPAE-PAD problems. Now, automated eluent generation is overcoming many of the issues and, together with increasingly effective chromatography columns, is improving the efficiency, reliability, and reproducibility of the analyses. Removing manual processes in glycan analysis workflows not only increases confidence in the critical results being generated, but it contributes to the increased productivity that laboratories are under increasing pressure to achieve.

References

1. H. Østby, et al., J. Chromatogr. A. 1662, 462691 (2022).
2. J. Rohrer, “Carbohydrate Analysis by High-Performance Anion-Exchange Chromatography with Pulsed Amperometric Detection (HPAE-PAD),” Thermo Fisher Scientific Technical Note, 2021.
3. J. Hu and J. Rohrer, “Determination of Glycoprotein Sialic Acid Composition Using HPAE-PAD in Dual Eluent Generation Cartridge Mode,” Thermo Fisher Scientific Application Note, 2021.
4. J.S. Rohrer, et al., Glycobiology, 26 (6) 582–591 (2016).
5. B. Huang and J. Rohrer, “HPAE-PAD Profiling of N-Linked Oligosaccharides from Glycoproteins Using Dual Eluent Generation Cartridges,” Thermo Fisher Scientific Application Note, 2019.

About the author

Wai-Chi Man is IC/SP product marketing manager at Thermo Fisher Scientific.

Article details

Pharmaceutical Technology
Volume 46, Number 11
November 2022
Pages: 36–41

Citation

When referring to this article, please cite it as W.C. Man, “Automating a Critical Step in Glycoprotein Analysis for Vaccine Manufacture," Pharmaceutical Technology 46 (11) 36–41 (2022).