Methylglyoxal in two acts: a genotoxic agent before transformation, and a flux-dependent survival filter after it

Methylglyoxal (MG) forms non-enzymatically from the triose phosphates of glycolysis at a rate first order in their concentration. It is therefore produced in proportion to glycolytic flux, and it is removed by a saturable enzymatic system. These two facts, taken together, make MG a quantity whose steady-state level is set by a ratio rather than by a rate. We argue that MG acts at two distinct points in tumour development, and that conflating them has obscured both. Before transformation, MG is genotoxic: it has been shown to transiently inactivate BRCA2 and to generate mutational signatures of defective homologous recombination without loss of the second allele, bypassing the two-hit requirement in heterozygous carriers. After transformation, once glycolytic flux has risen, MG becomes a constraint on survival: production scales with flux while clearance saturates, and a cell that raises flux without raising clearance capacity moves towards a boundary beyond which glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is progressively and irreversibly inactivated. We make the second claim quantitative. Writing the system as ordinary differential equations with separate pools for total and free MG, and calibrating against measured values - the formation constants of Phillips and Thornalley, the glyoxalase Michaelis constant measured in intact human erythrocytes, and the second-order rate constant for oxidation of the GAPDH active-site cysteine - the model places the transition at approximately 3.2 times basal glycolytic flux. The transition is sharp: between three- and fourfold flux, active GAPDH falls from 58 % to 5 %. The position of the threshold proves insensitive to the free MG fraction over a tenfold range, to the resting total MG pool over a threefold range, and to the absolute oxidation rate constant. It is sensitive to one quantity only: the fraction of GAPDH already irreversibly oxidised in a resting normal cell. That fraction is measurable by existing methods and has not, to our knowledge, been determined. We identify it as the single decisive experiment. We do not claim that MG initiates transformation. On the model's own logic it cannot: nothing in the MG-GAPDH-glyoxalase system generates the flux increase that the threshold requires. What the model describes is a filter acting on cells that have already acquired high flux by other means.