Two designer peptides pulled the accessory protein KCTD5 off the ion channel TRPM4, the channel stopped passing current, and triple-negative breast cancer cells stopped invading their environment. The route to the channel was not the channel.
The work, published online May 29 in the Journal of Chemical Information and Modeling ↗, is a structure-derived attempt to drug an ion channel sideways. TRPM4 is a calcium-activated cation channel whose currents drive cytoskeletal rearrangement, cell migration, and tumor cell invasion. It is elevated in breast tumors that do badly, and the field would like to inhibit it pharmacologically. The problem is that TRPM4 has been a poor small-molecule target. Pore blockers exist, but they are nonselective.
KCTD5 is a small protein that binds TRPM4 and increases its calcium sensitivity. Without KCTD5, TRPM4 still passes some current. With KCTD5 docked, it passes more. Knock KCTD5 out of breast cancer cells and the invasion phenotype softens. That made the TRPM4-KCTD5 interface itself an attractive target, on the logic that a peptide designed from one side of the interface should be able to occupy the contact patch and prevent the other side from docking.
What the group built and tested
The group at the Universidad de Chile designed two peptides in silico, each derived from one face of the interface. One reproduces a stretch of TRPM4 that contacts KCTD5; the other reproduces a stretch of KCTD5 that contacts TRPM4. Both carry a TAT cell-penetrating tag at one end and an HA epitope tag at the other, the standard format for delivering a peptide into a cell and being able to detect it later. They are referred to in the paper as TAT-TRPM4-HA and TAT-KCTD5-HA.
Three orthogonal assays. First, bimolecular fluorescence complementation, which fuses half of a fluorescent protein onto TRPM4 and the other half onto KCTD5, so that fluorescence lights up only when the two proteins are close enough to touch. Adding either peptide to HEK293 cells expressing the tagged pair dimmed the fluorescence. The interface was disrupted in the cell, not just on a docking screen.
Second, electrophysiology. Patch-clamp recordings and intracellular sodium measurements in HEK293 cells showed TRPM4-dependent currents drop after peptide treatment. Channel activity tracks the docking biochemistry.
Third, phenotype. MDA-MB-231 cells, a heavily used triple-negative breast cancer line, invade through a matrix in a standard transwell assay. The TAT-TRPM4-HA peptide reduced that invasion significantly compared with controls. The other peptide was characterized for binding and current but not explicitly for invasion in the abstract.
The cleanest control is genetic. The group ran their assays in parallel in HEK293 cells in which KCTD5 had been knocked out, which removes the protein the peptides are trying to displace. The peptides should have little to do in that background. That comparison is the built-in specificity test for whether the peptides really act on the KCTD5 axis or on something else entirely.
What this changes for someone working on peptides
Most ion channels are drugged at their pore. The pore is a tractable surface, well characterized, and has a long history of small-molecule blockers, but it is shared across channel subtypes in ways that limit selectivity. Going at the channel through its accessory protein is a different design philosophy. The interface between two proteins is large, irregular, and frequently undruggable by small molecules. Peptides, by contrast, are exactly the right size and shape to occupy a peptide-shaped contact patch. The TRPM4-KCTD5 example is one of a small but growing set of cases where this strategy has produced a working modulator.
The triple-negative breast cancer context is the harder part of the story to evaluate honestly. MDA-MB-231 cells in a transwell are not patients. The TAT-tag is a research delivery vehicle, not a drug formulation. The current data is a mechanism proof, not a candidate. But the design loop the paper closes (interface identified, peptides derived, biochemistry confirmed, channel function confirmed, invasion phenotype confirmed, knockout control available) is the loop that turns a target into a campaign. For ion channels that have resisted small-molecule drugging for decades, this is one of the more interesting moves in the toolkit.
There is one further structural point. The two peptides come from opposite sides of the same interface and both disrupt the complex. If only one had worked, you could blame it on a tag artefact or a coincidental side activity. Two peptides from two sides agreeing is the kind of internal consistency check that supports the mechanism as the primary explanation, not the only kind, but useful.