AI-designed miniproteins target 11 GPCRs

Researchers at the University of Washington’s Institute for Protein Design and Skape Bio said on May 21 that they had used AI-guided protein design to generate miniproteins that can activate or block G protein-coupled receptors, or GPCRs, a large class of drug targets. The results were published in Nature in a paper titled “De novo design of miniproteins targeting GPCRs.” The teams said the platform produced functional lead molecules against 11 GPCR targets. UW Medicine said the work could provide a new route to protein-based medicines for receptors involved in major disease pathways. ### Why are GPCRs such a big target in drug discovery? GPCRs are cell-surface receptors that help control many physiological processes, including senses such as sight and smell and signaling pathways used across the body. Because they sit in the cell membrane and regulate responses to hormones, neurotransmitters and other signals, they are among the most important target classes in medicine, according to UW Medicine and the Nature paper. (nature.com) Nature and UW Medicine said the challenge has been precision. Drug developers have long been able to hit GPCRs with small molecules or antibodies in some settings, but designing compact protein-based modulators that reliably switch specific receptors on or off has been harder. The new study describes a computational design approach aimed at that problem. ### What exactly did the researchers build? (nature.com) The Nature paper describes de novo designed miniproteins, meaning small proteins built computationally rather than copied from naturally occurring ones. UW Medicine said the designed proteins were made to bind GPCRs and act either as agonists, which activate receptors, or antagonists, which block them. Skape Bio said its platform combines AI-enabled protein design with high-throughput screening in living human cells. (nature.com) In its announcement, the company said the approach generated functional miniprotein agonists and antagonists against clinically important GPCR targets and that the study included structural data supporting how the designed proteins bind across multiple receptor families. ### How did they test whether the designs worked? Skape Bio said a central part of the work was a high-throughput screening system that tests designed proteins against full-length GPCRs in living human cells. The company said the system can evaluate up to 100,000 designs while keeping receptors in their natural membrane environment, where GPCR signaling occurs. (pharmiweb.com) The Nature abstract said the researchers used a “receptor diversion” microscopy-based screen to generate GPCR-binding miniproteins with high affinity and potency. That matters because GPCR behavior can change when receptors are removed from membranes or simplified for lab assays, so testing in cells is meant to preserve a more native setting. That last point is an inference from the study’s use of full-length receptors in living cells and Skape Bio’s description of the screening setup. (pharmiweb.com) ### What does “11 GPCR targets” actually tell us? The clearest concrete number in the announcement is 11. UW Medicine’s release and Skape Bio’s statement said the platform produced functional leads against 11 GPCR targets, indicating the method worked across multiple receptors rather than in a single proof-of-concept case. (nature.com) Skape Bio said its development focus includes migraine, metabolic disorders, inflammatory bowel disease, fibrosis and autoimmune conditions. The public materials tied the Nature paper to receptors involved in major disease pathways, but neither the UW release nor the search results reviewed here provided a full target-by-target efficacy summary suitable for comparing programs head-to-head. (newsroom.uw.edu) ### What happens next for this platform? The next step is drug development. Skape Bio said it is building GPCR biotherapeutics from the platform, while UW Medicine framed the study as a step toward medicines for diseases that still lack good treatments. Nature published the paper on May 21, and the company’s public pipeline page now lists programs in migraine, metabolic disease, IBD, fibrosis and autoimmune conditions. (skape.bio) Those programs will be the clearest place to watch whether the 11-target research result turns into named therapeutic candidates. (nature.com) (newsroom.uw.edu)