Quantitative profiling is essential for the identification of differential features in biomarker discovery. Among the technologies mentioned above, HPLC and CE are well known for their quantitative performance. The detection methods of glycans in these systems are based on the measurement of fluorescent signal of the derivatized glycans. As long as there is neither detector saturation nor substrate bias during derivatization, the signal intensity should be proportional to amount of analytes. CE has been demonstrated to be quantitative in analyzing native and derivatized glycosaminoglycans [68].
The quantitative aspect of glycan analysis in MALDI-TOF MS is affected by various parameters including analyte derivatization, types of laser, spotting protocol, choice of matrix, and acquisition protocol of spectrum. Structure of glycans, on the other hand, does not significantly affect the quantitative performance of MALDI-TOF MS [69], although it has been suspected that an extra antenna may promote postsource decay [61]. To obtain a quantitative profile of glycome by MALDI-TOF MS, methyl-esterification of the sialic acid residues is usually required to prevent the loss of sialic acid during MS analysis. However, our recent study has shown that sialylated glycan could be already quantified in the presence of 10 mM NaCl, but without the need of methyl-esterification [19]. In proteomic study, it had been shown that the reproducibility of MALDI-TOF mass spectra was greatly affected by the spotting methods, such as direct mixing, overlaying, and sandwiching [70]. Similarly, such parameter should also dictate the quantitative performance of MALDI-TOF MS in glycan analysis.
The reproducibility of the signal intensity of a glycan depends on (1) the quality of the glycan-matrix co-crystal on the sample spot and (2) the laser scanning protocol. The variation of signal intensities can be minimized by obtaining more mass spectra over different areas on the sample spot and averaging the resulting spectra. From a multi-institutional study, the intra-assay and inter-assay coefficients of variation percentages of glycans released from human IgG were below 10% and 4.2% for major species and 34% and 13% for minor species [61]. In a separate study, the intra-assay and inter-assay coefficients of variation percentages of three different standard glycans were below 9% and 18%, respectively [19]. This suggests that the quantitative performance of MALDI-TOF MS is comparable to chromatography. However, the multi-institutional study clearly illustrated that there were considerable variations between different centers [61]. More efforts have to be made to standardize the protocols and instrumentation to obtain comparable mass spectra from different research centers. The major shortcoming of MALDI-TOF MS is its incapability to provide concrete structure information of the glycan detected, although it is possible to predict structures of glycans based on molecular weights obtained from MS [19]. In the future, a similar approach could be carried out with a MALDI-TOF/TOF MS system to allow direct structure analysis by tandem MS. Besides a combined use of MALDI and TOF MS, MALDI has been combined with FTICR MS to obtain quantitative glycomic profiles [66, 67]. However, the reproducibility of the MALDI-FTICR MS in quantitative glycomic profiles has not been systematically evaluated.
Comparable to gene expression microarray and antibody microarray, lectin microarray was a newly developed technology for glycomics and glycoproteomics studies. Lectins are highly specific carbohydrate-binding biomolecules that recognizes glycans by their structural information. Different types of lectins recognize different oligosaccharide structures, terminal residue, and linkage [71]. For example, galectins are specific toward Gal residue [72]; siglecs are a group of I-type lectins, which mediate glycan recognition via Ig-like domains [73], and selectins are expressed by leukocytes and endothelial cells to recognize sialyl Lewis antigens [74]. Lectins have been widely used to study and isolate glycoproteins. Lens culinaris agglutinin lectin, which binds specific to alpha-1,6 fucosylation at proximal core GlcNAc, has been applied to isolate fucosylated AFP glycoforms associated with HCC [75]. Ricinus communis agglutinin lectin, which binds β-Gal, was applied on monitoring the desialylation of glycoproteins in murine B16 melanoma cells [76] and on investigating the degree of galactosylation of glycoproteins in human astrocytoma [77]. Sambucus nigra agglutinin lectin, which recognizes alpha 2,6-linked sialic acid, was used to measure the degree of alpha 2,6-sialylation of glycoproteins in human colon cancer [78]. Binding of Helix pomatia agglutinin lectin, which recognizes GalNAc, to human cutaneous malignant melanoma was found to be associated with metastasis formation [79]. Peanut lectin, which binds Galβ1−3GalNAc, was used to measure levels of mucin-type O-glycan in the human benign and malignant colorectal tissues [80].
In the lectin microarray technology, a panel of lectins is immobilized onto solid support in a microarray format. This allows simultaneous detection of glycans based on their interaction with different lectins. Lectin microarray was first developed in 2004 [81]. It has been shown that a lectin microarray spotted with nine lectins was already capable of giving distinct glycoprofiles for different glycoproteins [82]. In the experiment, lectin microarray was exposed to fluorescence-tagged glycoproteins, followed by washing and scanning. The experiment outcome was a pattern of lectins that interact with the glycans of glycoproteins. When a purified glycan or glycoprotein is analyzed with the lectin microarray, the structure of the glycan can then be interpreted based on the interaction pattern. Moreover, lectin microarrays can be used to study glycan–lectin interaction in a high-throughput manner [83].
Because the affinity of lectins to glycans was relatively low (Kd = 10–4 to 10–7M) compared with antibody–antigen interaction (Kd = 10–8 to 10–12M), various techniques have been developed to avoid loss of analytes during washing and to probe such weak interactions such as the evanescent-field fluorescence detection [83] and the ratiometric approach [84]. An evanescent field is a weak electric field that only propagates wavelength distance from the sensor surface. It allows real-time detection during the interaction event between lectins and glycans, thus omitting the need of washing. While the ratiometric approach utilizes dual fluorescence colors similar to DNA microarrays for different samples, allowing ratiometric comparison between glycomes. This approach was also shown to be able to analyze complex mixture including the total glycome of whole-cell lysates [84]. The quantitative performance of lectin microarray has been well evaluated [82]. The signal intensity was linear for glycoproteins in the range of 50 to 300 μg/ml. Coefficient of variation of signal intensities was within 10% to 20% [83].
Although lectin microarray also detects fluorescence signal associated with the glycans, there are several ways that are different from the approaches based on HPLC or CE. Firstly, in the lectin microarray technology, fluorescent tags could be attached to protein portions of the glycoconjugates, instead of the glycan molecules. Hence, the glycoproteins or glycoconjugates can be applied directly to the array without deglycosylation and desialylation. This simplifies the experimental process and allows more labeling options and downstream analysis as well as simultaneous analysis of O-linked, N-linked, and glycolipid glycans [82]. Secondly, lectin microarray belongs to the type of competition assays. Different glycoproteins having the same glycan structures will compete for the same group of lectins. It is not uncommon that different glycoproteins are labeled with a different number of fluorescent tags, leading to variations in the signal intensities between experiments. Furthermore, a single lectin molecule can bind to glycan molecules with different structures, but at different affinities. When a mixture of glycoproteins or glycoconjugates is subjected to a lectin microarray analysis, a highly complex pattern of competition and interaction among glycans and lectins will result.