Discovery, Design, and Evaluation of Nanomolar to Picomolar Inhibitors of Human Pancreatic α-Amylase for Control of Postprandial Blood Glucose Levels

Conspectus Type 2 diabetes is a disease that is rapidly expanding across the world with global cases quadrupling since 1990 and now comprising over 800 million people, the majority of whom are in low- and middle-income countries (Zhou et al.Lancet2024, 404, 2077–2093.). The poor control of blood glucose levels that lies at the heart of this disease leads to debilitating complications such as renal failure, vision loss, and amputations and has associations with heart disease and a host of other conditions. Oral drugs that control blood glucose levels and can slow disease progression and reduce the risks are available. Among these are the α-glucosidase inhibitors acarbose and miglitol that function in the gut to slow the breakdown of starch and of oligosaccharides such as sucrose. However, gastrointestinal side effects such as discomfort and diarrhea frequently cause patients to discontinue treatment, highlighting the need for improved, cost-effective alternatives. Our approach has been to develop selective inhibitors of human pancreatic α-amylase (HPA) that shut down digestion of starch while sparing oligosaccharide breakdown, thereby lowering blood glucose without causing gastrointestinal discomfort. In this Account, we present the approaches we have taken to identify potent and selective inhibitors of HPA based upon our earlier determinations of the three-dimensional structure and catalytic mechanism of this key digestive enzyme. Our screens of natural product extract libraries produced two particularly interesting hits. One is the natural product Montbretin A (MbA) isolated from the corms of the common Crocosmia garden flower. Through a series of fragmentation studies, in conjunction with X-ray crystallographic analysis of complexes of MbA with HPA, we identified the core inhibitory component of this (Ki = 8 nM) glycosylated acyl flavonoid comprising a pair of π-stacked phenolic moieties that engage the essential active site carboxylic acids in HPA. Animal studies have shown the efficacy and safety of MbA, and human clinical trials are underway. Based upon this information, simpler inhibitors that retain nanomolar potency have been designed and synthesized. Through a separate screen of marine natural product extracts, we identified helianthamide, from the Caribbean sea anemone Stichodactyla helianthus, as a particularly potent (Ki = 10 pM) and stable 44 amino acid peptidic inhibitor of HPA. The structure of its complex with the highly similar porcine pancreatic α-amylase (PPA) shows it blocking the active site, using two tyrosines and a histidine to engage the catalytic residues, reminiscent of the “phenol sandwich” found with MbA while deriving considerable additional affinity from the burial of surface area. Finally, we used mRNA display methods to screen large cyclic peptide libraries using the RaPID (Random nonstandard Peptide Integrated Discovery) system in collaboration with the Suga lab. Screening this cyclic peptide library, comprising ∼1012 peptides, yielded a 9-amino acid lariat peptide that is a potent (Ki = 7 nM) and selective inhibitor of HPA. Structural studies of the complex revealed once again a motif in which two tyrosines, separated by a proline, engage the active site residues. Thus, through screening of three completely distinct libraries, natural products from terrestrial and marine sources, as well as a large cyclic peptide library, a very similar inhibitory motif has been recovered in each case but from different classes of molecules–essentially a case of convergent inhibitor evolution.