October 15, 2012 (Vol. 32, No. 18)
Jeffrey Rohrer, Ph.D. Director of Applications Development Thermo Fisher Scientific
Deanna C. Hurum
A Comparison of Two Liquid Chromatographic Assay Methods
Sialic acids, also referred to as neuraminic acids, are critical to glycoprotein bioavailability, function, stability, and metabolism. When present, these carbohydrates occupy terminal positions of oligosaccharides in glycosylated proteins, providing charged points of interaction essential in many biological pathways.
Although over 50 natural sialic acids have been identified, two forms are commonly determined in therapeutic glycoprotein products: N-acetylneuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc). Due to loss of the enzyme CMP-N-acetylneuraminate monooxygenase, human tissue does not synthesize Neu5Gc, although it may be incorporated from dietary sources and has been studied as a disease marker.
Furthermore, Neu5Gc has been shown to be immunogenic, and the presence of this sialic acid in a therapeutic agent can potentially lead to an unwanted immune response. Therefore, total glycoprotein sialylation, and the identity of the sialic acids, play important roles in therapeutic protein efficacy, pharmacokinetics, and potential immunogenicity.
Glycoprotein sialylation analysis has been performed by many methods, ranging from colorimetric tests to sophisticated chromatographic methods. Chromatographic methods have the advantage of quantitatively differentiating sialic acids. The two chromatographic methods typically chosen for sialic acid determination are high-performance anion-exchange chromatography with pulsed amperometric detection (HPAE-PAD) and sample derivatization followed by UHPLC with fluorescence detection (UHPLC-FLD). These two methods each detect Neu5Ac and Neu5Gc, but by different approaches. HPAE-PAD relies on oxidation of the carbohydrate at a gold working electrode.
This detection method is direct and specific to the oxidation potential of the analyte. UHPLC-FLD relies on derivatization of the sialic acids to form a fluorescent product. Specificity is controlled by the derivatization reaction. Separation is subsequently performed on a C18 column after derivatization. This method is indirect and relies on the derivatization reaction for both analyte separation and detection.
For purposes of illustration, this article compares two independent chromatographic assays developed for sialic acids in five model glycoproteins: calf fetuin, bovine apo-transferrin (b. apo-transferrin), human transferrin (h. transferrin), sheep α1-acid glycoprotein (s. AGP), and human α1-acid glycoprotein (h. AGP). Analyses by both HPAE-PAD and UHPLC-FLD are evaluated with acid hydrolyzates of these proteins.
Fetuin (14 µg), h. transferrin (20 µg), b. transferrin (25 µg), h. AGP (13 µg), and s. AGP (7 µg) were added to individual 1.5 mL microcentrifuge vials with 200 µL of 2 M acetic acid. The protein solutions were hydrolyzed at 80°C for two hours. Two 50 µL aliquots of the hydrolyzate were taken. The first aliquot was prepared by lyophilization followed by dissolution in 500 µL of DI water before injection into the HPAE-PAD system. The second aliquot was derivatized with 1,2-diamino-4,5-methylenedioxybenzene dihydrochloride (DMB), a commonly used derivatization reagent for sialic acids.
An ion chromatography system (Thermo Scientific Dionex ICS-5000) with a Thermo Scientific Dionex CarboPac PA20 column was used for separation and subsequent detection by pulsed amperometry via a disposable gold working electrode and a four-potential waveform. A 10–200 mM sodium acetate in 100 mM NaOH gradient from 0–15 min followed by 5 min at 200 mM sodium acetate in 100 mM NaOH at a flow rate of 0.5 mL/min were used to elute sialic acids. The column was equilibrated at 10 mM sodium acetate in 100 mM NaOH for 5 min before a 10 µL sample injection. Both the column and detector cell were maintained at 30°C.
Derivatized sialic acids were separated on a UHPLC system (Thermo Scientific Dionex Ultimate 3000 RSLC). Separation was performed on a Thermo Scientific Acclaim RSLC 120 C18, 2.2 µm, 2.1 x 100 mm at a flow rate of 0.45 mL/min and a temperature of 45°C under isocratic conditions with 9:7:84 acetonitrile:methanol:water mobile phase. Samples were maintained at 4°C in the dark in the autosampler. Analytes were detected by fluorescence (373 nm λex, 448 λem).
Sialic acids analysis is possible by both HPAE-PAD and UHPLC-FLD methods. Figure 1A shows the separation of sialic acids in protein hydrolyzates and a standard by HPAE-PAD. Neu5Ac elutes first, and both Neu5Ac and Neu5Gc are easily determined within 10 min. Figure 1B shows the separation of sialic acids from the same protein hydrolyzates after they have been derivatized with DMB.
In this method Neu5Gc elutes first and the identified sialic acids are eluted within 5 min; however, depending on the samples, the runtime may need to be extended to 10 min to avoid interferences from other components. In both cases, linear calibration is possible, with coefficients of determination for Neu5Ac of >0.9995 between 1–100 pmol (HPAE-PAD) and >0.995 between 2.1–50 pmol (UHPLC-FLD).
The Table compares the determined values for the five glycoproteins evaluated. Due to inherent variability in acid hydrolysis, analysis precision both within one day and across several days can be highly variable by either chromatographic assay. Results can vary within 20% from one hydrolysis to the next, however; with the exception of fetuin, precision for triplicate analysis were within 10%. Results for fetuin ranged between 18–22% (RSD). This variability is dependent on the sialic acid release and not on the chromatographic method.
Acid hydrolysis is a balance between sialic acid release from the glycoprotein versus degradation of the released carbohydrates. For this reason, optimization of this step for each sample is important to control analysis variability. Additionally, for UHPLC-FLD analysis, the derivatization reaction must also be optimized for the sample being analyzed, and the standard curve must be prepared in a matrix that matches the samples.
HPAE-PAD is a direct method that does not require derivatization. Both HPAE-PAD and UHPLC-FLD can be used to determine Neu5Ac and Neu5Gc in proteins with similar results, with the caveat that the alkaline elution conditions used in HPAE-PAD can deacetylate the oxygen-linked acetyl groups of O-acetylated sialic acids converting them to Neu5Ac and Neu5Gc.
If determination of individual O-acetylated sialic acids is needed, the UHPLC-FLD method is preferred. However, UHPLC-FLD requires sample derivatization, which requires additional time for routine analysis of Neu5Ac and Neu5Gc. Both methods require the same sample hydrolysis optimization for consistent sample analysis. In addition to the sample hydrolysis, derivatization will require 2.5 h for the reaction and additional time to stop the reaction and prepare the samples for injection after the derivatization is complete. In this comparison a 150 mm column was used for HPAE-PAD, but analysis time can be reduced with equivalent results by using a 30 mm format column.