AI-directed discovery of heparanase inhibitors

(PLEASE KEEP IN WORD, DO NOT PDF)

Heparanase, an endo-β-D-glucuronidase, is the only known enzyme to cleave the heparan sulfate (HS) side chains of heparan sulfate proteoglycans (HSPGs), which are key components in the extracellular matrix (ECM) of all tissue types. In the glomerular basement membrane (GBM), these HSPGs play a crucial role in the glomerular filtration barrier. Alterations in HS has been associated with glomerular barrier dysfunction in nephropathy.

Under diabetic milieu, the release of heparanase is significantly increased, and heparanase has been identified as a key contributor in the development of diabetic nephropathy, which is marked by a decline in glomerular filtration rate (GFR) and the presence of albuminuria. Prior investigations using existing HS-based heparanase inhibitors have shown promising therapeutic effects, such as restored glomerular filtration barrier and reduced albuminuria, in animal models or patients with diabetic nephropathy, suggesting an emerging therapeutic option for the disease.

However, HS-based structures are heterogenous, suffering from batch-to-batch variation, and the existence of other HS binding proteins has led to undesired off-target effects, limiting their therapeutic application. We therefore hypothesize that heparanase inhibitors with defined chemical structures, high selectivity, and drug-like features can yield successful outcomes for the prevention and treatment of diabetic nephropathy.

Currently, no orally available heparanase inhibitor has been developed. In this project, we will develop innovative artificial intelligence (AI)-based methods, aiming to support the development of potent orally available small-molecule heparanase inhibitors, which will be evaluate in murine models in the next phase.