Formation of methylglyoxal-adducts in plasma is associated with LDL and triglyceride level: implications for diabetic atherosclerosis

Aims Protein glycation leading to AGE is enhanced in diabetes by increases in blood glucose per se, and collaterally, by endogenous production of reactive carbonyls. Low weight -dicarbonyls are formed as glycolytic intermediates during metabolic conversion of glucose and/or during lipid peroxidation. Among AGE precursors, methylglyoxal (MG) is considered as one of the key intermediates. We hypothesized it to be a common product of both carbonyl and oxidative stress, and investigated the MG biogenesis in relation to glycemic and lipid status in diabetic patients. Methods Serum and urine MG-adduct content was measured by DELFIA method in 83 diabetic patients and 20 controls. Fasting (FG) and postprandial (PPG) glucose level, HbA1c, LDL and HDL cholesterol, plasma triglyceride and homocysteine level were determined along with routine biochemical parameters. Results Significant positive relationship was observed between serum level of MG-adducts and LDL (r=0.31 p=0.003), whereas FG, PPG and HbA1c correlated inversely (r= -0.33 p=0.001 ; r= -0.23 p=0.041 ; r= -0.22 p=0.036). Similarly, significant correlations were also found between urinary levels of MG-adducts and PPG (r= -0.28 p=0.023), serum triglycerides (r=0.31 p=0.003), homocysteine (r=0.57 p=0.0007), HDL (r= -0.28 p=0.007) and urine albumin/creatinine ratio (r=0.53 p=0.002). Stepwise linear regression was performed using serum or urine MG-adducts as a dependent variable and HbA1c, fasting and postprandial glucose, LDL, HDL, triglycerides, serum creatinine, homocysteine and urine albumin/creatinine ratio as independent variables. Of these, only LDL-cholesterol (regression coefficient=0.29) and FG (regression coefficient=-0.28) were independent predictors of MG-adducts in serum (p 3.0 mmol/L discriminated patients who had a higher serum level of MG-adducts (median (10th and 90th percentile) 465 (251-1254) vs 331(169-706) mgEq/L, p=0.0052), although there was no between-subgroup difference in glycemic control. Patients on statin treatment had lower MG-adducts although the difference did not reach statistical significance. The positive relationship between LDL-c and MG-adducts (r=0.38, p=0.042) was noted in the patients (n=54) free of statin treatment, whereas in the statin-treated subgroup, there was an inverse but not significant tendency (r=-0.28, p=0.83). A significant correlation between homocysteine and urinary excretion of MG-adducts (r=0.8 ; p=0.02) was recorded in patients with a history of macrovascular disease. Conclusion A highly significant relationship between LDL and MG-adduct production, as well as tight correlation between triglycerides and urinary MG-adduct excretion suggest that lipoxidation and glyceraldehyde-3-phosphate route, along with the glycolytic pathway, might be an important source of MG generation. The glycotoxin methylglyoxal seems to be a common factor linking the two dominant metabolic changes in diabetes, hyperglycemia and intensive lipolysis, with the vascular pathobiochemistry of diabetes.