Design-rules for stapled peptides with in vivo activity and their application to Mdm2/X antagonists
Stapled peptides reach targets small molecules cannot, but the changes that get them into cells are the same ones that make them toxic. This work turns that trade-off into explicit design rules — then proves them by applying them to a different series.
A staple holds the helix — and, this work suggests, does more than that: the choice of crosslink also shaped how selective the molecule turned out to be.
Most of the proteome is not druggable by conventional means. Antibodies cannot cross the membrane. Small molecules need a deep hydrophobic pocket to grip, which transcription factors and most intracellular protein–protein interfaces do not provide. Stapled α-helical peptides were proposed as the answer to exactly this gap: large enough to cover a flat interface, constrained into a helix by a synthetic crosslink, and — in principle — able to get inside a cell.
In practice the field has struggled with two problems it has never fully separated. Making a stapled peptide permeable usually means making it more hydrophobic or more positively charged, and both of those changes tend to bring toxicity that has nothing to do with the intended target. The result is a molecule that looks active in cells for the wrong reason.
A consortium spanning MSD International Singapore, Merck & Co., A*STAR, Uppsala University and EPOC Scientific set out to replace intuition with rules. They synthesised more than 350 molecules against Mdm2 and MdmX — the negative regulators of p53 — and published, in Nature Communications, what amounts to a worked manual.
Five findings, stated as rules
The paper’s contribution is that its conclusions are actionable rather than descriptive.
Lipophilicity tracks permeability, clearly and measurably — but the same axis runs toward insolubility and off-target trouble, so it cannot simply be maximised.
Positive charge causes off-target toxicity. Peptides carrying multiple positive charges frequently produced lactate dehydrogenase release — membrane damage — and non-specific effects. Removing that charge removed the problem.
Anionic residues, placed judiciously, buy solubility and better behaviour — but placement matters, not merely count. A glutamate in the wrong position costs permeability; in the right position it rescues a peptide’s handling without doing so.
The C-terminus is a potency lever. Extending it with polyalanine improved cellular activity, with a sweet spot around a six-alanine amide tail; shorter and longer tails were both slightly worse in the standard assay, though longer extensions performed better under serum.
Staple chemistry governs polypharmacology. This is the least intuitive result and the most useful.
The staple is not just a clamp
Replacing the conventional olefin staple with a triazole crosslink preserved high-affinity binding — 1.1 nM — but hurt cellular potency badly, to 13.6 µM in the presence of serum. Adding a second staple as an N-terminal lactam bridge gave moderate potency but poor solubility, apparently because it removed a negative charge.
The rigid di-alkyne staple was the one that worked. Combined with a polyalanine tail, it gave good solubility (128 µM) and good activity: 480 nM in the ten percent serum p53 reporter assay, and 124 nM in an HCT116 proliferation assay.
More importantly, both the double-stapled and the di-alkyne peptides were remarkably free of off-target toxicity across the cell proliferation panel — which distinguished them not only from the other stapled peptides tested but from advanced Mdm2 small-molecule antagonists MK-4688 and AMG 232.
The authors do not claim to know why. They note the mechanism is not clear, and offer as speculation that it may relate to reduced promiscuous binding, through restricting the conformational freedom of the staples or the backbone. It is an honest gap in an otherwise prescriptive paper, and it points at something the field has not resolved: a staple is usually treated as a means of enforcing helicity, and this result suggests it is also determining selectivity.
Applying the rules
Run together, the workflow produced peptides with more than 292-fold improved cell proliferation potency over the starting point and no off-target proliferation effects — an on-target index above 3800-fold.
The stronger test was transfer. The team applied the same rules to a structurally distinct Mdm2(X) peptide series, and improved its cellular potency by more than 150-fold while removing its off-target toxicities. Rules derived from one scaffold worked on another, which is the difference between a set of observations and a method.
Into an animal
The in vivo comparison was run in a mouse SJSA-1 osteosarcoma xenograft, dosed intravenously every third day at 30 mg/kg. ATSP-7041 — the widely studied template for this class — produced clear tumour growth inhibition of about 33%. Both ALRN-6924, the clinical molecule also known as sulanemadlin, and the di-alkyne stapled peptide MP-616 outperformed it substantially, each achieving around 66% tumour growth inhibition.
Although not examined here, it would be interesting to understand if the clean off-target profile for MP-616 translates into fewer side-effects in vivo.
That caveat is the authors’ own, and it matters. The cleanliness of MP-616 was established in cell panels. Whether it produces a better safety profile in an animal or a patient was not tested, and the paper does not assert it.
What this is, and is not
This is a methods paper wearing the clothes of a medicinal chemistry campaign. Its value is not the individual molecules but the fact that more than 350 of them were made and measured consistently enough to yield rules that transferred to a different series.
The scope is bounded. The design rules are derived from, and validated on, Mdm2(X) antagonists — targets with hydrophobic interfaces addressable by a helical motif, which is the class the authors explicitly say the workflow should suit. How much generalises to targets with different interface chemistry is not established here. The in vivo work is a single xenograft model at one dose and schedule. And the mechanism behind the most interesting finding — that staple choice governs off-target behaviour — remains open.
What the work does supply is unusual in this field: a set of stated, tested, transferable rules for a modality that has too often advanced by trial and error, together with the control experiments needed to tell genuine cellular activity from membrane damage.
Source
Arun Chandramohan, Hubert Josien, Tsz Ying Yuen, Ruchia Duggal, Diana Spiegelberg, Lin Yan, Yu-Chi Angela Juang, Lan Ge, Pietro G. Aronica, Hung Yi Kristal Kaan, Yee Hwee Lim, Andrea Peier, Brad Sherborne, Jerome Hochman, Songnian Lin, Kaustav Biswas, Marika Nestor, Chandra S. Verma, David P. Lane, Tomi K. Sawyer, Robert Garbaccio, Brian Henry, Srinivasaraghavan Kannan, Christopher J. Brown, Charles W. Johannes and Anthony W. Partridge. “Design-rules for stapled peptides with in vivo activity and their application to Mdm2/X antagonists.” Nature Communications, 2024, volume 15, article 489. Received 1 April 2023; accepted 6 November 2023; published online 12 January 2024. DOI: 10.1038/s41467-023-43346-4.
Affiliations: MSD International, Singapore; Merck & Co., Inc., Kenilworth, New Jersey, Boston, Massachusetts and West Point, Pennsylvania; Institute of Sustainability for Chemicals, Energy and Environment, A*STAR, Singapore; Bioinformatics Institute, A*STAR, Singapore; Institute of Molecular and Cell Biology, Singapore; Department of Surgical Sciences and Department of Immunology, Genetics and Pathology, Uppsala University, Sweden; EPOC Scientific LLC, Stoneham, Massachusetts. Corresponding authors: Brian Henry, Srinivasaraghavan Kannan, Christopher J. Brown, Charles W. Johannes and Anthony W. Partridge.
The underlying research is the work of the cited authors and their institutions. Maestro Chronicle provides the editorial presentation.