Vilya designs new macrocyclic peptides

Throw a rock in today’s biotechnology landscape, and it will probably hit an artificial intelligence drug discovery firm. Throw that rock in Seattle, and it’ll probably hit an AI protein design firm spun off from newly minted Nobel laureate David Baker’s lab at the Institute for Protein Design (IPD).

One of those Baker lab spin-offs is Vilya, which uses computational protein design to create membrane- permeable macrocyclic drugs. When the biotech venture capital firm Arch Venture Partners met the Vilya team in December 2021, its partners needed just 5 min to be convinced of Vilya’s potential because of how advanced its discovery process already was.

“Once we saw how the technology was actually able to make molecules that were both potent and permeable, the data really spoke for itself,” says Inca Dieterich, a partner at Arch who became one of Vilya’s cofounders and is its vice president of business development.

Arch was also struck by just how much of Vilya’s work could be done in silico, Dieterich says. While a lot of drug discovery companies go through massive brute-force screens of synthesized molecules, Vilya is able to do much of its screening in the computational realm, allowing it to focus on the most promising molecules in the wet lab.

IPD faculty member Gaurav Bhardwaj, IPD researchers Adam Moyer and Nao Hiranuma, IPD senior fellow Patrick Salveson, IPD chief strategy and operations officer Lance Stewart, and Baker cofounded the firm with Arch on the basis of science published in the journal Cell (2022, DOI: 10.1016/j.cell.2022.07.019). Vilya formally launched in August 2022 with Moyer, Hiranuma, and Salveson on staff.

In contrast with linear peptides, which are lines of amino acids, macrocyclic peptides have amino acids arranged in a ring. Vilya CEO Cyrus Harmon says a ring is typically more desirable because linear peptides are “floppy” and tend to get metabolized quickly.

“We have much more rigid, well-defined shapes in the molecules that we’re making,” Harmon says, adding that this rigidity helps with accurately predicting the peptide’s structure and binding ability.

Vilya is also designing new lab-made, noncanonical amino acids and incorporating them into protein designs, expanding the range of peptides it can create.

“If you think about the 20 naturally occurring amino acids, 19 of them differ only in the side chain,” Harmon says. “We can incorporate lots of interesting heterocycles. We can incorporate very exotic side chains.”

In April, Vilya researchers published a paper showing that they had used their computational platform to design 15 million potential macrocycles using amino acids beyond the 20 natural ones (Science 2024, DOI: 10.1126/science.adk1687). Chemists then synthesized 18 of these macrocycles, 15 of which were “very close to the design models.” Around the same time, the company expanded its series A fundraising round, which raised a total of $71 million.

Macrocycles could be the sweet spot between small molecules and larger biologics, Dieterich says. Small molecules struggle to target a large protein surface, but it’s difficult to make biologics that can be administered orally.

“We can tackle things that are intracellular, that disrupt protein-protein interactions, and that’s what makes certain macrocycles really advantageous therapeutics when we’re looking across the whole space of various therapeutic approaches,” Dieterich says.

Harmon told C&EN in October that the company is developing candidates in the gastroenterological, oncological, and immunologic fields.