Fig 1.
Model NHA cells and experimental design for the glycomic analysis.
(a) Generation of immortalized and transformed human astrocytes. (b) Expression levels of specific defined elements (i.e., hTERT, SV40ER (both large T and small t), H-RasV12, and myrAKT) assessed by immunoblotting. (c) Protocols employed for glycomic analyses of N-, O- and GSL-glycans. For further details, see Methods.
Fig 2.
Quantified GSL-glycans were stratified into ganglio (Gg), globo (Gb), (neo)lacto (n)Lc, “Gg or (n)Lc”, and “Gb or (n)Lc” groups according to glycan structure (a). Glycans classified into the Gg (b), Gb (c) and (n)Lc-series (d) were further compared according to the levels of constitutive GSL-glycans. SphinGOMAP (http://www.sphingomap.org/) online databases were used for structural estimation of GSL glycans, which are summarized in S1 Table. Biosynthetic pathways for the generation of Gg-, Gb-, and (n)Lc-series GSL-glycans (e) and real-time PCR analysis of glycosyltransferases involved in GSL-glycan biosynthesis (f). Each value represents the mean ± SD of three independent real-time PCR analyses. Comparison of the relative abundance of neutral and sialylated GSL-glycans (g), and comparison of relative levels of GSL-glycans with N-glycolylneuraminic acid incorporation among total sialylated GSL-glycans (h).
Fig 3.
Expression profiles of representative GSL-glycans.
(a) Gg-series glycans. (b) Gb-series glycnas. (c) MALDI-TOF/TOF MS spectra of GSL-glycans with a (Hex)5(HexNAc)2 composition. In theory, signature fragmentation peaks at m/z 934.9 and 975.4 are specifically generated from Gb- and (n)Lc-series GSL-glycans, respectively. (d) Expression profiles of Gb-series (Hex)5(HexNAc)2 glycan (left) and (n)Lc-series (Hex)5(HexNAc)2 glycans (left). The glycan amounts in (d) were calculated from the area of the signature fragmentation peaks (m/z 934.9 and 975.4) shown in (c). SphinGOMAP (http://www.sphingomap.org/) online databases were used for the structural estimation of the GSL glycans. Each value represents the mean ± the standard deviation (SD) of three independent MALDI-TOF MS-derived analyses.
Fig 4.
Quantified N-glycans were structurally classified as PM-, HM-, and C/H-type glycans (a). PM- (b), HM- (c), C/H neutral- (d), and C/H acidic-type glycans (e) were further compared according to composition. Glycan classifications were performed based on the estimated glycan structures and compositions shown in S6–S10 Tables. Fucosylation and sialylation status of C/H-type N-glycans (f and g). (h) Proportion of α2,6 sialic acids among total sialic acids with A2 and A1 (left) and A2F and A1F (right). (i) Relative levels of N-glycans showing N-glycolylneuraminic acid incorporation among total sialylated N-glycans. Expression profiles of glycans with bisecting GlcNAc residues (j) and tetra-antennary glycans (k). (j) Sum of the expression levels of all glycans having (Hex)2(HexNAc)3 or (Hex)3(HexNAc)4 substructures were compared among NHA, NHA/T, NHA/TS, NHA/TSR, and NHA/TSRA cells. (k) Sum of the expression levels of all glycans having (Hex)4(HexNAc)4 as a substructure was compared among the different cell types. Left; absolute amounts, right: relative abundance of tetra-antennary glycans among C/H-type glycans. (l) Real-time PCR analysis of MGAT3 (left) and MGAT5 (right) expression levels. Each value represents the mean ± the standard deviation (SD) of three independent MALDI-TOF MS (f, g, j, and k) or real-time PCR analyses (l).
Fig 5.
Quantified O-glycans were classified according to their composition (a), number of sialic acid (Sia) residues (b), and core glycan structures (c). Comparison of relative levels of O-glycans showing N-glycolylneuraminic acid (Neu5Gc) incorporation among total sialylated O-glycans (d). Estimated biosynthetic pathway of O-glycans in NHA cells (e). Glycan classifications were based on the estimated glycan structures and compositions shown in S3 Table. Each value shown in b and d represents the mean ± range of two independent MALDI-TOF MS analyses.
Fig 6.
Classification of cells and glycans by N-, O- and GSL-glycomics based on unsupervised cluster analysis.
The absolute amount of each glycan (pmol/100 μg protein) was analyzed by using Cluster 3.0 software. The heat map with clustering was acquired by using Java Threeview software. The relative abundance of each glycan (classified into Group A, B-1, B-2, B-3, C, D or E) is shown in the lower panel.
Fig 7.
Representative glycans expressed at high levels upon ectopic expression of the various genes.
Table 1.
Summary of the causal relationship between multistep tumorigenesis and glycomic alterations.