Researchers built a short peptide, reinforced its shape with a molecular staple, and used it to jam a transcription switch that a hard-to-treat lung cancer depends on. The stapled peptide bound its target at low-nanomolar strength, slipped into cells without tearing their membranes, and choked the growth of cancer cells missing a tumor suppressor called LKB1, while barely touching cells that still had it. All of that happened in a dish. No animal has seen the drug yet.
The target nobody has drugged well
Roughly a fifth of non-small-cell lung cancers have lost LKB1, a gene that normally keeps cell growth in check by switching on a metabolic brake called AMPK. Tumors without it grow aggressively and shrug off most current treatments, including immunotherapy. The problem is that LKB1 is a tumor suppressor, something the cancer has deleted, and you cannot drug an absence. So the target has to be whatever the loss switches on.
A group led by Lizi Wu at the University of Florida had earlier traced that switch to a transcription program called CRTC-CREB. When LKB1 is gone, CRTC and CREB, two proteins that dock together to turn genes on, run unchecked and drive the malignancy. CREB is a classic hard target. It works through a broad, flat protein-to-protein contact, the kind of interface that small-molecule drugs, built to plug deep pockets, tend to slide right off.
Stapling the helix so it holds its shape
The team, reported in Molecular Therapy: Oncology ↗, took the stretch of CRTC that grips CREB and rebuilt it as a standalone peptide. On its own, such a fragment flops into a floppy coil and loses the spiral shape it needs to bind. The fix was a hydrocarbon staple, a small chemical brace welded across two turns of the helix to lock it rigid. The approach comes from the Dana-Farber lab of Loren Walensky, a co-author and one of the people who developed stapled peptides in the first place.
The braced versions, which the authors call SAH-CBDs (stabilized alpha helices of the CREB binding domain), held a stable spiral, bound directly to CREB, and blocked CRTC and CREB from forming their complex in the test tube. They made six variants and settled on one, SAH-CBD-6, as the lead. Its binding sat in the low-nanomolar range (strong, meaning it grabs its target at tiny concentrations), and in a competition assay the stapled peptides beat an unstapled control by five to ten times. Crucially, the peptides crossed into cells without the membrane damage that many cell-penetrating peptides cause, then shut down CREB-driven gene activity.
Selective in a dish
The payoff was selectivity. In LKB1-null lung cancer lines (A549, H157, H1437), SAH-CBD-6 suppressed growth. In cells that still carried working LKB1 (H322, and normal BEAS-2B bronchial cells), it did far less. That is the pattern you want from a targeted drug: it hurts the cells defined by the vulnerability and mostly spares the rest.
The reasons for caution are equally clear, and the authors list them. This is cell-culture work. There is no mouse, no tumor implant, no dosing in a living animal. The paper itself flags pharmacokinetics, distribution, tumor penetration, and safety as future work, which is another way of saying none of it has been tested. Selectivity in a dish is not selectivity in a body, where a peptide has to survive the bloodstream, reach the tumor, and leave healthy tissue alone. Transcription factors like CREB have a long record of looking druggable in vitro and staying stubborn in vivo.
What the work does establish is narrower and still useful. A stapled peptide can reach CREB, hold the shape it needs, get into cells cleanly, and act on the exact cells that lost LKB1. It is a prototype and a tool, not a therapy. The same logic, a braced peptide reaching a driver that small molecules cannot grip, has shown up against other so-called undruggable targets, including a peptide that jammed the cancer driver FOXM1 ↗ and an oral peptide aimed at mutant RAS ↗. Whether SAH-CBD-6 joins them in an animal is the next test, and the one this paper does not yet run.