Heterozygosity for a loss-of-function mutation in GALNT2 improves plasma triglyceride clearance in man

Abstract

Genome-wide association studies have identified GALNT2 as a candidate gene in lipid metabolism, but it is not known how the encoded enzyme ppGalNAc-T2, which contributes to the initiation of mucin-type O-linked glycosylation, mediates this effect. In two probands with elevated plasma high-density lipoprotein cholesterol and reduced triglycerides, we identified a mutation in GALNT2. It is shown that carriers have improved postprandial triglyceride clearance, which is likely attributable to attenuated glycosylation of apolipoprotein (apo) C-III, as observed in their plasma. This protein inhibits lipoprotein lipase (LPL), which hydrolyses plasma triglycerides. We show that an apoC-III-based peptide is a substrate for ppGalNAc-T2 while its glycosylation by the mutant enzyme is impaired. In addition, neuraminidase treatment of apoC-III which removes the sialic acids from its glycan chain decreases its potential to inhibit LPL. Combined, these data suggest that ppGalNAc-T2 can affect lipid metabolism through apoC-III glycosylation, thereby establishing GALNT2 as a lipid-modifying gene.

Figure 1. 2DE Analysis of ApoC-III Isoforms in Carriers of the GALNT2^D314A Mutation Compared to Noncarriers and Enzyme Kinetics of Wild-Type and Mutant ppGalNAc-T2

(A) Compared to controls, carriers have 6.6-fold increased levels of nonsialylated apoC-III₀ (p = 0.01) and decreased levels of monosialylated apoC-III (apoC-III₁; p = 0.05), while levels of disialyated apoC-III (apoC-III₂) are similar. Data are expressed as means. Error bars are SEMs. P values are for Mann-Whitney U tests. (B) Western blot of 2DE gel showing a carrier with increased levels of apoC-III₀ isoforms, while apoC-III of a family control is only present as apoC-III₁ or apoC-III₂. Note that apoC-III₀ comprises nonsialylated apoC-III (apoC-III-GalNAc and apoC-III-GalNAc-Gal) and apoC-III that is not glycosylated (Bruneel et al., 2007). (C and D) Enzymes were overexpressed in COS7 cells. y axis indicates rate of transfer as dpm/hr/densitometric unit of enzyme. Figures represent kinetic plots and averages of replicate peptide glycosylation experiments catalyzed by wild-type or mutant ppGalNAc-T2. Mixtures of purified recombinant proteins with various concentrations of EA2 (C) or apoC-III (D) peptide substrates were used, and kinetic plots were fit to Michaelis-Menten equation. Kinetic parameters for peptide glycosylation by both enzymes are given in tables as means. Error bars are SD. N, number of replicate experiments.

Figure 2. Carriers of the GALNT2^D314A Mutation Have Improved Postprandial Plasma Triglyceride Clearance

Depicted are plasma triglyceride levels before (t = 0) and after an oral fat challenge (t = 1–6 hr) in four carriers and four noncarriers. Compared to noncarriers (open symbols), carriers (closed symbols) have lower baseline triglyceride levels, earlier maximum triglyceride levels (t = 3 versus t = 4 in controls), and a reduced overall triglyceride increase. The estimated difference in plasma triglycerides (averaged over time) between carriers and noncarriers using a mixed linear model was 15.4 mg/dl (95% confidence interval 3.3–27.5; p = 0.014). Data are expressed as means. Error bars are SEM. The inset shows that the oral fat challenge did not change plasma LPL levels in carriers and controls when using a commercially available ELISA. The LPL levels appear lower in carriers compared to controls, but these were not significantly different from controls at any time point as assessed by unpaired Student’s t tests. Data are given as means. Error bars are SD.

Figure 3. Inhibition of Lipoprotein Lipase-Mediated Triglyceride Hydrolysis

(A) 2DE analysis of untreated purified human apoC-III that was isolated from plasma VLDL (Academy Bio-Medical Company). In addition to the common apoC-III₂ and apoC-III₁ isoforms, apoC-III in VLDL is also present in a more acidic isoform(denoted apoC-III₃), as has been described previously (Wopereis et al., 2003; Jabs and Assmann, 1987). The second panel shows that treatment of this apoC-III with neuraminidase results in desialylation of apoC-III, as evidenced by the appearance of the nonsialylated isoform, apoC-III₀. (B) Desialylation of this apoC-III with neuraminidase attenuates its potential to inhibit recombinant human LPL. LPL activity is inhibited by 54% after incubation with untreated apoC-III. Desialylation with neuraminidase reduces the inhibitory capacity of apoC-III (p < 0.001). Vertically shaded bar indicates that catalytic activity of LPL is not affected by neuraminidase treatment. Experiments were conducted in triplicate. Data are expressed as means. Error bars are SD.