Fig. 5: GAPDH is inversely correlated with methylglyoxal production, GLO1 abundance and redox status.

A An open search of LFQ proteomics data detected a lower quantity of MG-driven protein modifications (MG-H1 on arginine and CEL on lysine) in tumor compared to healthy adjacent tissue (N = 70 patients, healthy or tumor tissue, two-sided paired Student’s t test). B A targeted search for MG-H1 modification corroborated the open search output (N = 70 patients, healthy or tumor tissue, two-sided paired Student’s t test). C GAPDH activity can control G3P and DHAP levels, which are precursors of MG. Correspondingly, D tumor MG-H1 modification frequency (as % of all PSMs), s-lactoylglutathione (sLG) values (as tumor/healthy (T/H) ratio) and tumor GLO1, but not GLO2 protein levels, inversely correlate with the abundance of GAPDH. Numbers of pairs per correlation analysis (i.e., number of patients): 69 (GLO2) and 70 (all other). E Kaplan–Meier plots of PANCAN dataset (10-year survival) for GAPDH (N (number of patients) = 11,506). F GAPDH abundance was positively correlated with fluorodeoxyglucose (FDG) uptake of tumors and (G) with GLO1 oxidation levels (N pairs for correlation analysis (i.e., number of patients): 63 or 59 correspondingly).