Phosphomannose isomerase inhibitors improve N-glycosylation in selected phosphomannomutase-deficient fibroblasts

Abstract

Congenital disorders of glycosylation (CDG) are rare genetic disorders due to impaired glycosylation. The patients with subtypes CDG-Ia and CDG-Ib have mutations in the genes encoding phosphomannomutase 2 (PMM2) and phosphomannose isomerase (MPI or PMI), respectively. PMM2 (mannose 6-phosphate → mannose 1-phosphate) and MPI (mannose 6-phosphate ⇔ fructose 6-phosphate) deficiencies reduce the metabolic flux of mannose 6-phosphate (Man-6-P) into glycosylation, resulting in unoccupied N-glycosylation sites. Both PMM2 and MPI compete for the same substrate, Man-6-P. Daily mannose doses reverse most of the symptoms of MPI-deficient CDG-Ib patients. However, CDG-Ia patients do not benefit from mannose supplementation because >95% Man-6-P is catabolized by MPI. We hypothesized that inhibiting MPI enzymatic activity would provide more Man-6-P for glycosylation and possibly benefit CDG-Ia patients with residual PMM2 activity. Here we show that MLS0315771, a potent MPI inhibitor from the benzoisothiazolone series, diverts Man-6-P toward glycosylation in various cell lines including fibroblasts from CDG-Ia patients and improves N-glycosylation. Finally, we show that MLS0315771 increases mannose metabolic flux toward glycosylation in zebrafish embryos.

FIGURE 1.

Source of mannose in N-glycans depends on PMM2/MPI ratio. Nine cell lines were double-labeled with [6-³H]glucose and [4-¹⁴C]mannose at the same specific activity in the medium containing physiological concentration of 5 mm glucose and 50 μm mannose. The proportion of mannose-derived and glucose-derived mannose in N-glycans was calculated based on mannose analysis of acid hydrolysate and plotted against the ratio of enzymatic activities (PMM2/MPI). Cell lines assayed were CDG-Ib patient fibroblasts, CRL1947 fibroblasts, 42F fibroblasts, human umbilical vein endothelial cells, SKNSH neuroblastoma cells, C2C12 muscle cells, hepatocytes, C3a hepatoma cells, and 293 kidney cells.

FIGURE 2.

siRNA-mediated MPI inhibition increases mannose flux for glycosylation. HT-29 human colonic epithelial cells were treated with three MPI siRNAs at different concentrations to block MPI activity to varying levels followed by labeling with [2-³H]mannose and glycoprotein precipitation using TCA. a, MPI activity in knockdown cells as compared with untreated cells. b, [2-³H]mannose incorporated in glycoproteins in siRNA-treated cells was plotted against the percentage of MPI inhibition (calculated using data from panel a).

FIGURE 3.

MLS0315771 is a potent MPI inhibitor in cells. a, HeLa cells were labeled in the absence and presence of different concentrations of MLS0315771 with 50 μCi/ml [2-³H]mannose and 5 μCi/ml [³⁵S]Met/Cys. The proteins were precipitated with TCA, and ³H incorporation in N-glycans was normalized to ³⁵S incorporation. Data are an average of three independent determinations. Results with inhibitor were compared with those without inhibitor to calculate p values (Student's t test) *, p = 0.02–0.05, **, p = 0.01 or less, ns = not significant. b, inhibitor efficacy was compared in HeLa cells, C2C12 muscle cells, 42F fibroblasts, C3a hepatoma cells, and HT-29 colon cells. The cell lines were labeled with 50 μCi/ml [2-³H]mannose in the presence of different concentrations of MLS0315771, and [2-³H]mannose in the precipitated glycoproteins was determined. c, HeLa cells were labeled with 50 μCi/ml [2-³H]mannose in the absence and presence of different concentrations of MLS0315771. The cells were harvested and lysed at the indicated times (0, 30, 60, 90, 120 min). [2-³H]Mannose incorporation in the precipitated glycoproteins was determined.

FIGURE 4.

MLS0315771 toxicity at higher concentrations is an off-target effect. a, HeLa cells were labeled with [³⁵S]Met/Cys for 3 h in the presence of MLS0315771. ³⁵S incorporation in TCA-precipitated glycoproteins was determined. The data with inhibitor were compared with data without inhibitor to calculate p values. *, p = 0.02–0.05, **, p = 0.01 or less, ***, p = 0.001 or less, ns = not significant. Significance values were calculated as compared with untreated control. b, mouse embryonic fibroblasts from WT and Mpi-KO mice were labeled with [³⁵S]Met/Cys, and ³⁵S incorporated into proteins was determined. Data presented are an average of three independent determinations.

FIGURE 5.

MLS0315771 increases mannose flux toward glycosylation. a and b, normal and CDG-Ia fibroblasts were preincubated with MLS0315771 for 2 h followed by labeling with 50 μCi/ml [2-³H]mannose and 5 μCi/ml [³⁵S]Met/Cys for 1 h in the presence or absence of different concentrations of mannose without (a) or with MPI inhibitor (b). ³H and ³⁵S incorporation was determined in precipitated glycoproteins. Data presented here are representative of two independent determinations. c and d, normal (c) and CDG-Ia (d) fibroblasts were preincubated with MLS0315771 for 1 h and then labeled for 4 h with stable labels: 5 mm [1,2-¹³C]glucose and 50 or 200 μm [4-¹³C]mannose. The glycans were isolated and hydrolyzed, and mannose was analyzed by GC/MS. The percentage of total N-glycan labeled with each isotope was determined. Numbers on top of each bar indicate the relative contribution to mannose from each isotope. Error bars indicate a range between two fragments used to the calculate abundance of each sugar (m242 and m314 for [1,2-¹³C]glucose origin and m217 and m314 for [4-¹³C]mannose origin). Data presented here are representative of two independent determinations.

FIGURE 6.

A subset of CDG-Ia patient fibroblasts responds to MLS0315771 treatment. Samples were treated with Endo H before precipitation and compared with untreated sample. a, Endo H sensitivity of CDG-Ia fibroblast line with or without inhibitor in the presence of mannose. b and c, Endo H sensitivity of different CDG-Ia fibroblasts bearing different mutations in PMM2 labeled in the absence (b) or presence (c) of the indicated amount of mannose. d, stable isotope labeling of three CDG-Ia patient fibroblasts with 5 mm [1,2-¹³C]glucose and 50 or 200 μm [4-¹³C]mannose and GC/MS analysis of mannose in N-glycans from CDG-Ia fibroblasts. − indicates no inhibitor, and + indicates 10 μm MLS0315771. Details of this analysis are same as in Fig. 5d. Data presented here are representative of two independent determinations.

FIGURE 7.

MLS0315771 modulates mannose flux in zebrafish embryos. a, 2-day old zebrafish embryos (10 per well) were treated without or with different concentrations of the inhibitor, and survival was determined over 2 days. Results are from two independent experiments. b, 4-day-old zebrafish embryos (30 per well) were labeled with 50 μCi/ml [2-³H]mannose and 5 μCi/ml [³⁵S]Met/Cys in the absence or presence of the indicated amount of MPI inhibitor. Radiolabel was measured in glycoproteins precipitated from whole zebrafish lysate. Data are an average of two independent experiments. Significance values were calculated as compared with untreated control. *, p = 0.02–0.05, **, p = 0.01 or less, ns = not significant.