Sorbitol Is a Severity Biomarker for PMM2-CDG with Therapeutic Implications

Abstract

Objective: Epalrestat, an aldose reductase inhibitor increases phosphomannomutase (PMM) enzyme activity in a PMM2-congenital disorders of glycosylation (CDG) worm model. Epalrestat also decreases sorbitol level in diabetic neuropathy. We evaluated the genetic, biochemical, and clinical characteristics, including the Nijmegen Progression CDG Rating Scale (NPCRS), urine polyol levels and fibroblast glycoproteomics in patients with PMM2-CDG.

Methods: We performed PMM enzyme measurements, multiplexed proteomics, and glycoproteomics in PMM2-deficient fibroblasts before and after epalrestat treatment. Safety and efficacy of 0.8 mg/kg/day oral epalrestat were studied in a child with PMM2-CDG for 12 months.

Results: PMM enzyme activity increased post-epalrestat treatment. Compared with controls, 24% of glycopeptides had reduced abundance in PMM2-deficient fibroblasts, 46% of which improved upon treatment. Total protein N-glycosylation improved upon epalrestat treatment bringing overall glycosylation toward the control fibroblasts' glycosylation profile. Sorbitol levels were increased in the urine of 74% of patients with PMM2-CDG and correlated with the presence of peripheral neuropathy, and CDG severity rating scale. In the child with PMM2-CDG on epalrestat treatment, ataxia scores improved together with significant growth improvement. Urinary sorbitol levels nearly normalized in 3 months and blood transferrin glycosylation normalized in 6 months.

Interpretation: Epalrestat improved PMM enzyme activity, N-glycosylation, and glycosylation biomarkers in vitro. Leveraging cellular glycoproteome assessment, we provided a systems-level view of treatment efficacy and discovered potential novel biosignatures of therapy response. Epalrestat was well-tolerated and led to significant clinical improvements in the first pediatric patient with PMM2-CDG treated with epalrestat. We also propose urinary sorbitol as a novel biomarker for disease severity and treatment response in future clinical trials in PMM2-CDG. ANN NEUROL 20219999:n/a-n/a.

FIGURE 1:

Epalrestat treatment increased phosphomannomutase (PMM) enzyme activity and ICAM-1 protein abundance. (A) Epalrestat treatment increased PMM enzyme activity in patients with PMM2-CDG “fibroblasts” (n = 11; P1–P6, P8, P10, P17, P19, and P24). The graphs represent results after 10 μM epalrestat treatment for 24 hours (P5 responded to the dose of 5 μM epalrestat with 35% increase, P6 did not respond to any of the doses with enzyme activity increase; patient deceased; Table S2). (B) Quantification of ICAM-1 protein abundance in immunoblots with patients with PMM2-CDG “fibroblasts” based on band intensity (p = 0.02; n = 10; P1–P6, P8, P17, P19, and P24). (C) Quantification of immunoblots showing LAMP-2 protein abundance in epalrestat untreated and treated patients’ “fibroblasts” (n = 10; P1–P6, P8, P17, P19, and P24). Epalrestat treatment does not increase LAMP-2 protein abundance. (D) Immunoblots showing ICAM-1 protein abundance in epalrestat untreated and treated patients’ “fibroblasts” (n = 10; P1–P6, P8, P17, P19, and P24). Beta-Actin was used as a loading control. (E) Immunoblots showing LAMP-2 protein abundance in epalrestat untreated and treated patients’ fibroblasts (n = 10; P1–P6, P8, P17, P19, and P24). Beta-Actin was used as a loading control.

FIGURE 2:

Proteomic changes in phosphomannomutase (PMM)-deficient patient derived fibroblasts and effect of epalrestat treatment. Panel A depicts the waterfall plot of global proteomics of PMM2-congenital disorders of glycosylation (CDG) and controls. Y-axis is log₂ fold changes (PMM2-CDG/controls) and X-axis is number of proteins identified. Each individual circle represents a protein. Some of the highly changing representative protein names are marked by triangles and proteins names are provided. (B) Volcano plot for the same comparison as panel A is shown. X-axis is log₂ fold change (PMM2-CDG/controls) and Y-axis is the negative logarithm of p value of t test for significance. The horizontal dashed line marks the cutoff for significance (<0.05) and vertical dashed lines are drawn to highlight the proteins having at least 30% change in either direction (1.3-fold enhancement or reduction). Some of the most significantly changing proteins are marked by triangles and proteins names are provided. (C) Waterfall plot is depicted for paired comparison of PMM-deficient fibroblasts treated with epalrestat or untreated (vehicle). Y-axis is log₂ fold changes (epalrestat-treated/untreated) and X-axis is number of proteins identified. Each identified protein is depicted with a black circle and some of the highly changing proteins are marked by triangles and proteins names are provided. (D) Volcano plot is depicted for treated/untreated comparison of PMM2-deficient fibroblasts. X-axis depicts log₂ fold change (treated/untreated) and Y-axis is the negative logarithm of p value of t test for significance. Horizontal dashed line marks the cutoff for significance (paired t test, <0.05) and the vertical dashed lines are drawn to highlight the proteins having at least 30% change in either direction (1.3-fold enhancement or reduction). Using this cutoff, some of the proteins showing relative higher abundance upon epalrestat treatment are shown as triangles.

FIGURE 3:

Glycoproteome alterations in phosphomannomutase (PMM)-deficient fibroblasts and remodeling upon epalrestat treatment. (A) The waterfall plot of global glycoproteomics of PMM-deficient fibroblasts and controls. Y-axis is log₂ fold changes (PMM2-congenital disorders of glycosylation [CDG]/controls) and X-axis is number of unique glycopeptides identified. Each individual triangle represents a unique glycopeptide (unique combination of peptide and glycan structure). Some of the highly changing representative glycopeptides are marked by triangles and glycoprotein names, glycosylation site (N with corresponding amino acid number) and plausible glycan structures are marked. The oval in the lower half (negative Y-axis) of the waterfall plot depicts unique glycopeptides having the plausible glycan structures (shown in the box above the oval), which were reduced in PMM-deficient fibroblasts. (B) Volcano plot for the given comparison (PMM2-CDG/controls) is shown. X-axis is log₂ fold change (PMM2-CDG/controls) and Y-axis is the negative logarithm of p value of t test for significance. The horizontal dashed line marks the cutoff for significance (<0.05) and the vertical dashed lines are drawn to highlight the glycoproteins names having at least 30% change in either direction (1.3-fold enhancement or reduction). Some of the highly changing glycopeptides are marked by triangles and glycoproteins’ names, glycosylation sites and plausible glycan structures are drawn. (C) Waterfall plot is depicted for paired comparative glycoproteomics of PMM-deficient fibroblasts treated with epalrestat or untreated (vehicle). Y-axis is log₂ fold changes (epalrestat-treated/untreated) and X-axis is number of unique glycopeptides identified and quantified. Each unique glycopeptide is depicted with a black circle and some of the altered glycopeptides are marked by triangles and glycoproteins’ names, glycosylation sites and plausible glycan structures are marked. (D) Volcano plot is depicted for treated/untreated comparative glycoproteomics of PMM-deficient fibroblasts. X-axis depicts log₂ fold change (treated/untreated) and Y-axis is the negative logarithm of p value of t test for significance. The horizontal dashed line marks the cutoff for significance (paired t test, <0.05) and the vertical dashed lines are drawn to highlight the glycopeptides having at least 30% change in either direction (1.3-fold enhancement or reduction). Using this cutoff, some of the glycopeptides showing enhanced levels upon epalrestat treatment are shown as triangles. With this cutoff none of the unique glycopeptides was found to be reduced.

FIGURE 4:

Waterfall plot of 2 different comparisons at the protein level. (A) This panel depicts the waterfall plot of global proteomics of phosphomannomutase-congenital disorders of glycosylation (PMM2-CDG) and controls. Y-axis is log₂ fold changes (PMM2-CDG/controls) and X-axis is number of proteins identified. Each individual circle represents a protein name. Some of the highly changing representative proteins are marked by triangles and proteins names are provided. (B) Waterfall plot is depicted for paired comparison of PMM-deficient fibroblasts treated with epalrestat or untreated (vehicle). Y-axis is log₂ fold changes (epalrestat-treated/untreated) and X-axis is number of proteins identified. Each identified protein is depicted with a black circle and some of the highly changing proteins are marked by triangles and proteins names are provided. These waterfall plots are shown side-by-side for comparison.

FIGURE 5:

Glycopeptides with the highest improvement (%) after epalrestat treatment. Epalrestat treated (solid line) and untreated (dashed line) phosphomannomutase-congenital disorders of glycosylation (PMM2-CDG) patient-derived fibroblasts were compared with controls and 10 glycopeptides showed the highest percent of improvement in their relative abundance levels are shown in this figure. Every datapoint is one glycopeptide and their corresponding protein name, glycosylation site, and plausible glycan structure are marked at the bottom.

FIGURE 6:

Urine sorbitol and mannitol concentrations in patients with peripheral neuropathy, liver pathology, and congenital disorders of glycosylation (CDG) phenotype. Concentrations were normalized to urine creatine concentration. Significant variation in urine sorbitol and mannitol concentrations were associated with both peripheral neuropathy score and liver pathology score, with elevated urine sorbitol and mannitol detected in patients with CDG displaying both moderate neuropathy and mild liver pathology. The Kruskal-Wallis test followed by Dunn’s multiple comparisons test (A, C) and the Mann Whitney test (B, D). Data are expressed as mean ± SD. p < 0.05(*), p < 0.01(**), and p < 0.001(**). Urine sorbitol levels positively correlated with severe CDG phenotype (r = 0.5, p < 0.02) (E). There was no correlation with the mild or moderate category. Urine mannitol levels did not correlate with mild, moderate, or severe categories based on Nijmegen Progression CDG Rating Scale (NPCRS) scores (F).

FIGURE 7:

Epalrestat has a positive effect on the glycosylation defect, growth, and normalization of elevated sorbitol and mannitol levels in the phosphomannomutase-congenital disorders of glycosylation (PMM2-CDG) pediatric patient. (A) Body mass index (BMI) of the patient increased substantially following epalrestat treatment. (B) Pharmacokinetics profile shows rapid elimination of epalrestat. The samples representing each time point were taken from different dosing days. Epalrestat was eliminated with a terminal half-life (t½) of approximately 1 to 2 hours. Blood samples were provided were drawn: prior to therapy (0 hour) and 1 hour post-dose (month 6), 1 hour 20 minutes (day 1), 1 hour 40 minutes (day 90), 2 hours (month 9), 3 hours (day 30), 4 hours (day 2), 6 hours (month 12), and 8 hours 30 minutes (day epalrestat was eliminated with a t½ of approximately 1 to 2 hours. A peak concentration of epalrestat in blood was 1125.408 ng/ml (or 3.5 μM) 1 hour after the epalrestat administration. A trough level as the lowest concentration reached by epalrestat before the next dose was 23.392 ng/ml after 8 hours. P1 single-dose pharmacokinetic (PK) data on 0.27 mg/kg epalrestat dose three times. (C) Urine sorbitol level before starting epalrestat therapy (sorbitol = 19.93 mmol/mol creatinine), after 3 months (5.78 mmol/mol creatinine) and 6 months of therapy (6.20 mmol/mol creatinine) compared with controls (<5 mmol/mol creatinine). (D) Urine mannitol level before starting epalrestat therapy (mannitol = 648.6 mmol/mol creatinine), after 3 months (37.32 mmol/mol creatinine) and 6 months of therapy (25.09 mmol/mol creatinine) compared with controls (<20 mmol/mol creatinine). (E) Increasing weight during epalrestat therapy and decreasing blood transferrin glycoform ratio analysis 12 months prior to therapy and during 12 months of epalrestat therapy. Normal level for Mono-oligo/Di-oligo Controls: Ratio ≤0.06 and A-oligo/Di-oligo Controls: Ratio ≤0.01

FIGURE 8:

Hypothetical mechanism of altered polyol metabolism in phosphomannomutase-congenital disorders of glycosylation (PMM2-CDG). Decrease in PMM enzyme activity is associated with an excess of metabolites including sorbitol and mannitol.