GLUCOSE and other reducing sugars react with proteins by a non-enzymatic, post-translational modification process called non-enzymatic glycosylation or glycation. The sugar-derived carbonyl group adds to a free amine, forming a reversible adduct which over time rearranges to produce a class of products termed advanced-glycation end-products (AGEs). These remain irreversibly bound to macromolecules and can covalently crosslink proximate amino groups^1,2. The formation of AGEs on long-lived connective tissue and matrix components accounts largely for the increase in collagen crosslinking that accompanies normal ageing and which occurs at an accelerated rate in diabetes^3,4. AGEs can activate cellular receptors and initiate a variety of pathophysiological responses^5–9. They modify an appreciable fraction of circulating low-density lipoproteins preventing uptake of these particles by their high-affinity tissue receptors^10,11. Advanced glycation has also been implicated in the pathology of Alzheimer's disease^12,13. Because AGEs may form by a pathway involving reactive α-dicarbonyl intermediates^1,2,14, we investigated a potential pharmacological strategy for selectively cleaving the resultant glucose-derived protein crosslinks. We now describe a prototypic AGE crosslink 'breaker', N-phenacylthiazolium bromide (PTB), which reacts with and cleaves covalent, AGE-derived protein crosslinks. The ability of PTB to break AGE crosslinks in vivo points to the importance of an α-dicarbonyl intermediate in the advanced glycation pathway and offers a potential therapeutic approach for the removal of established AGE crosslinks.