Glucoselysine, a unique advanced glycation end-product of the polyol pathway and its association with vascular complications in type 2 diabetes

Abstract

Glucoselysine (GL) is an unique advanced glycation end-product derived from fructose. The main source of fructose in vivo is the polyol pathway, and an increase in its activity leads to diabetic complications. Here, we aimed to demonstrate that GL can serve as an indicator of the polyol pathway activity. Additionally, we propose a novel approach for detecting GL in peripheral blood samples using liquid chromatography-tandem mass spectrometry and evaluate its clinical usefulness. We successfully circumvent interference from fructoselysine, which shares the same molecular weight as GL, by performing ultrafiltration and hydrolysis without reduction, successfully generating adequate peaks for quantification in serum. Furthermore, using immortalized aldose reductase KO mouse Schwann cells, we demonstrate that GL reflects the downstream activity of the polyol pathway and that GL produced intracellularly is released into the extracellular space. Clinical studies reveal that GL levels in patients with type 2 diabetes are significantly higher than those in healthy participants, while N^δ-(5-hydro-5-methyl-4-imidazolon-2-yl)ornithine (MG-H1) levels are significantly lower. Both GL and MG-H1 show higher values among patients with vascular complications; however, GL varies more markedly than MG-H1 as well as hemoglobin A1c, fasting plasma glucose, and estimated glomerular filtration rate. Furthermore, GL remains consistently stable under various existing drug treatments for type 2 diabetes, whereas MG-H1 is impacted. To the best of our knowledge, we provide important insights in predicting diabetic complications caused by enhanced polyol pathway activity via assessment of GL levels in peripheral blood samples from patients.

Keywords: advanced glycation end-products (AGEs); aldose reductase; biomarker; blood; diabetes; fructose; glucose metabolism; glucoselysine; hyperglycemia; mass spectrometry (MS); polyol pathway.

Figure 1

Measuring of GL levels in human serum.A, GL and/or [¹³C₆] GL were mixed with or without FL and/or [¹³C₆] FL, and subjected to acid hydrolysis. The peak area of the GL (open bar) and [¹³C₆] GL (striped bar) were measured using LC-MS/MS. The values in the graphs represent the mean ± SD of at least three independent experiments (n = 3). Dunnett test. B, the peak area of GL in the serum of healthy participants and patients with type 2 diabetes changed with (open bar) and without (striped bar) acid hydrolysis. The values in the graphs represent the mean ± SD of four healthy participants, and four patients with type 2 diabetes. Individual data points were represented by closed circles. HP, healthy participants; T2D, patients with type 2 diabetes. The significance between hydrolysis and no hydrolysis is indicated with asterisk (paired t test), and the significance between healthy participants and patients with type 2 diabetes is indicated with dagger (Welch's two-sample t test). ∗p < 0.05, ∗∗p < 0.01, ^†p < 0.05. C, quantification of GL in the serum of healthy participants. The values in the graphs represent the mean ± SD of at least three independent experiments (n = 3). One-way ANOVA with Bonferroni correction, ∗∗p < 0.01, ∗∗∗p < 0.001. D, representative chromatograms of GL and [¹³C₆] GL in the serum of healthy participants, obtained by LC-MS/MS. FL, fructoselysine; GL, glucoselysine; LC-MS/MS, liquid chromatography-tandem mass spectrometry.

Figure 2

Relationship between GL and fructose levels and the degree of changes in their levels in human serum.A, relationship between GL and fructose levels in human serum. r_s, Spearman rank correlation coefficient. B, degree of changes in GL and fructose levels in patients with type 2 diabetes (T2D) was calculated relative to their levels in healthy participants (HP), serving as the reference. The values in the graphs represent mean ± SD of at least three independent experiments (n = 4). Individual data points were represented by closed circles. The significant difference between healthy participants and patients with T2D is indicated with a dagger (Mann–Whitney U test). ^†p < 0.05. GL, glucoselysine.

Figure 3

Intracellular and extracellular GL contents under high glucose and fructose conditions. Levels of GL in intracellular (A) and extracellular (B) compartments of WT (open bar) and immortalized aldose-reductase-KO (ARKO, striped bar) Schwann cells 72 h after exposure to normal (5.5 mM) and high glucose (50.5 mM) conditions. The values in the graphs represent the mean ± SD of at least three independent experiments (n = 3). Individual data points were represented by closed circles. The statistical significance of the difference between WT and ARKO is indicated with an asterisk, and the statistical significance of the difference between normal and high glucose conditions is indicated with a dagger. One-way ANOVA with Bonferroni correction, ∗∗p < 0.01, ^††p < 0.01, ^†††p < 0.001. Levels of GL in intracellular (C) and extracellular (D) compartments of WT (open bar) immortalized Schwann cells 72 h after exposure to 0, 10, and 100 μM fructose. The values in the graphs represent the mean ± SD of at least three independent experiments (n = 3). Individual data points were represented by closed circles. One-way ANOVA with Bonferroni correction, ∗p < 0.05, ∗∗p < 0.01. GL, glucoselysine.

Figure 4

Comparison of GL and MG-H1 levels between healthy participants and patients with type 2 diabetes. Amounts of GL (A), MG-H1 (B), Lys (C), and Arg (D) in the serum of healthy participants (HP, open box, n = 21) and patients with type 2 diabetes (T2D, dense dotted box, n = 153). The concentration of the internal standard in the human serum was 10 pmol. Welch's two-sample t test, ∗p < 0.05, ∗∗p < 0.01. GL, glucoselysine; MG-H1, N^δ-(5-hydro-5-methyl-4-imidazolon-2-yl)ornithine.

Figure 5

Changes in the serum GL, MG-H1, and HbA1c levels depending on vascular complications. Levels of GL (A), MG-H1 (B), and HbA1c (C) in patients with type 2 diabetes and cooccurrent microvascular complications (MICRO). Patients with type 2 diabetes without (open box, n = 60) or with (dense dotted box, n = 93) MICRO. Levels of GL (D), MG-H1 (E), and HbA1c (F) in patients with type 2 diabetes and cooccurrent macrovascular complications (MACRO). Patients with type 2 diabetes without (open box, n = 115) or with (dense dotted box, n = 38) MACRO. Levels of GL (G), MG-H1 (H), and HbA1c (I) levels in patients with type 2 diabetes and cooccurrent micro and/or macrovascular complications (VCOMP). Patients with type 2 diabetes without (open box, n = 51) or with (dense dotted box, n = 102) VCOMP. The concentration of the internal standard in the human serum was 10 pmol. Welch's two-sample t test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. GL, glucoselysine; HbA1c, hemoglobin A1c; MG-H1, N^δ-(5-hydro-5-methyl-4-imidazolon-2-yl)ornithine.

Figure 6

Summary of fluctuations in glucose metabolism-related AGEs in type 2 diabetes.Dashed arrows represent promotion, while T arrows indicate inhibition. AGE, advanced glycation end-product; DHAP, dihydroxyacetone phosphate; eGFR, estimated glomerular filtration rate; G3P, glyceraldehyde 3-phosphate; GLP-1, glucagon-like peptide-1; glucose-6P, glucose 6-phosphate; GLUT, glucose transporter.