Mass spectrometric quantification of plasma glycosphingolipids in human GM3 ganglioside deficiency

Abstract

Background: Among Amish communities of North America, biallelic mutations of ST3GAL5 (c.694C > T) eliminate synthesis of GM3 and its derivative downstream a- and b-series gangliosides. Systemic ganglioside deficiency is associated with infantile onset psychomotor retardation, slow brain growth, intractable epilepsy, deafness, and cortical visual impairment. We developed a robust quantitative assay to simultaneously characterize glycan and ceramide moieties of plasma glycosphingolipids (GSLs) among ST3GAL5 c.694C > T homozygotes (n = 8), their heterozygous siblings (n = 24), and wild type control (n = 19) individuals.

Methods: Following extraction and saponification of total plasma lipids, GSLs were purified on a tC18 cartridge column, permethylated, and subjected to nanospray ionization mass spectrometry utilizing neutral loss scanning and data-dependent acquisition. Plasma GSLs were quantified against appropriate synthetic standards.

Results: Our method demonstrated linearity from 5 to 250 μl of plasma. Recovery of synthetic GSLs spiked into plasma was 99-104% with no matrix interference. Quantitative plasma GSL profiles discriminated among ST3GAL5 genotypes: GM3 and GD3 were undetectable in ST3GAL5 c.694C > T homozygotes, who had markedly elevated lactosylceramide (19.17 ± 4.20 nmol/ml) relative to heterozygous siblings (9.62 ± 2.46 nmol/ml) and wild type controls (6.55 ± 2.16 nmol/ml). Children with systemic ganglioside deficiency had a distinctive shift in ceramide composition toward higher mass species.

Conclusions: Our quantitative glycolipidomics method discriminates among ST3GAL5 c.694C > T genotypes, can reveal subtle structural heterogeneity, and represents a useful new strategy to diagnose and monitor GSL disorders in humans.

Keywords: CID, collision-induced dissociation; Cer, ceramide; Dp, degree of polymerization; EGCase, endoglycosylceramidase; ESI-MS, electrospray ionization mass spectrometry; GD3, disialo-ganglioside GD3 (IUPAC-IUB: II3- α -(Neu5Ac)2-Gg2Cer); GM1b, monosialo-ganglioside GM1b (IUPAC-IUB: IV3-α-Neu5Ac-Gg4Cer); GM3; GM3, monosialo-ganglioside GM3; GSL, glycosphingolipid; Gal, galactose; GalNAc, N-acetylgalactosamine; Ganglioside; Gb3, globotriaosylceramide (IUPAC-IUB: Gb3Cer); Gb3-D, deuterated Gb3; Gb4, globotetraosylceramide (IUPAC-IUB: Gb4Cer); Glc, glucose; GlcCer, glucosylceramide; Glycosphingolipid; LacCer, lactosylceramide; MS, mass spectrometry; MSn, multidimensional mass spectrometry; Mass spectrometry; NL, neutral loss; NSI, nanospray ionization; Neu5Ac, sialic acid as N-5-acetylneuraminic acid; Plasma; ST3GAL5, CMP-Neu5Ac:Lactosylceramide alpha-2,3-sialyltransferase 5, previously known as SIAT9, SIATGM3S, ST3GalV, GM3-synthase; TIM, total ion mapping; UPLC, ultra-high pressure liquid chromatography.

Graphical abstract

Fig. 1

Biosynthetic pathways for glycosphingolipid production. Lactosylceramide (LacCer) is produced by the glycosylation of ceramide (Cer) with glucose (Glc) and subsequent elongation with galactose (Gal). LacCer is a substrate for the production of multiple glycosphingolipid (GSL) types, including lacto- and neolacto-series, by extension with N-acetylglucosamine (GlcNAc), globo-series, by extension with Gal, or ganglio-series, by extension with N-acetylgalactosamine (GalNAc). Addition of sialic acid (Neu5Ac) to LacCer, catalyzed by ST3GAL5, generates GM3, the precursor for the production of all complex a- and b-series gangliosides. In the absence of ST3GAL5, the complex a- and b-series gangliosides are lost (red box) and LacCer is shunted toward production of externally sialylated o-series gangliosides as well as globo- and lacto-/neolacto-series GSLs.

Fig. 2

NSI-MS profiles of plasma GSLs. Representative full MS spectra are shown for analysis of GSLs prepared from plasma obtained from wild type control (+/+), ST3GAL5 c.694C > T heterozygous siblings, (M/+) and ST3GAL5 c.694C > T homozygous subjects (M/M). Each GSL glycan is associated with multiple peaks that reflect the heterogeneity of the ceramide portion of the molecule. Mass peaks attributed to GM3 are shaded in pink. Dp4, the maltotetraose calibrant peak is shaded red and was added in equal amount to all three samples.

Fig. 3

Quantification of plasma GSLs. GSLs were quantified in plasma obtained from wild type controls (+/+, n = 24), ST3GAL5 c.694C > T heterozygous siblings (M/+, n = 19), and ST3GAL5 c.694C > T homozygous subjects (M/M, n = 8). (A) Box-and-whisker plots present concentration ranges, means, and quartiles for GlcCer (1), LacCer (2), Gb3 (3), Gb4 (4), GM3 (5), GD3 (6), GM1b (7). (B) Direct comparative analysis of wild type, heterozygous, and homozygous subjects for LacCer, GM3, GD3, and GM1b levels. (C) Precursor/product ratios for the indicated lipids. ns, not significant; ^**P < 0.01, ^****P < 0.0001.

Fig. 4

Plasma LacCer ceramide heterogeneity. (A) Representative zoom spectra for the mass range containing LacCer ceramide forms are presented for a wild type control (+/+) and an ST3GAL5 c.694C > T homozygous subject (M/M). The m/z values at which LacCer species were detected are shown along the x-axis. (B) Summary of the mean relative abundances of LacCer ceramide forms in wild type controls and ST3GAL5 c.694C > T homozygous subjects. Pie slices are color-coded to match the m/z values in panel A.

Fig. 5

Heterogeneity of plasma GSL ceramides. For each of the indicated GSLs, MS peak intensities associated with the most abundant ceramides bearing short (c16-c20, black pie slices) or long (c22-24, grey pie slices) N-acyl chains were summed together. The summed intensities for short or long N-acyl chains were normalized to the total intensity and are plotted as pie charts to compare the relative abundance of ceramide forms across wild type control (+/+), ST3GAL5 c.694C > T heterozygotes (M/+), and homozygotes (M/M). ns, not significant; ^****P < 0.0001; nd, not detected; *P < 0.05.