A loop of RNA that most cells treat as noise turns out to encode a small protein, and that protein helps lung tumors shrug off one of their most important drugs.
The drug is osimertinib, sold as Tagrisso by AstraZeneca, a pill that blocks mutated EGFR and is the standard first treatment for non-small-cell lung cancer driven by the common EGFR mutations, exon 19 deletions and the L858R substitution. It even handles T790M, the mutation that used to defeat earlier EGFR drugs. It works well until it stops. Tumors reliably learn to resist it, and why they do is only partly understood. Writing online August 18 in Cellular Signalling ↗, a group building osimertinib-resistant lung cancer cell lines went looking for the difference between cells that still respond and cells that no longer do.
They found a circular RNA they named circTLL1, sharply elevated in the resistant cells. Circular RNAs are loops with no free ends, long filed under noncoding "junk" and studied mostly as sponges that soak up regulatory microRNAs. This one was doing something else. Forcing circTLL1 up made sensitive cells resistant. Knocking it down made resistant cells respond to osimertinib again, both in the dish and in tumors grown in mice.
The reason is a hidden reading frame. Buried inside circTLL1 is a stretch that cells actually translate into a 90-amino-acid protein, which the authors call circTLL1-90aa. Peptides this small, encoded where nobody expected a gene, are a fast-growing category of biology, and most of them still have no known job. This one has a clear one.
What the small protein does
circTLL1-90aa latches onto a metabolic enzyme called NT5C2 and marks it for destruction. NT5C2 normally keeps a lid on the cell's nucleotide supply. With it degraded, GTP, one of those nucleotides and also a fuel for growth signaling, builds up, and the buildup keeps the Ras and PI3K/AKT pathways switched on. Those are the same growth circuits osimertinib is trying to quiet by blocking EGFR further upstream. The small protein gives the tumor a way to keep the downstream engine running even after the upstream switch is turned off.
That makes circTLL1-90aa two things at once, the authors argue: a marker that could flag which tumors are slipping toward resistance, and a target a future drug could aim at to keep osimertinib working longer.
Both claims are early. Everything here is cell lines and mouse xenografts. There is no patient data in the report, and a mouse tumor is a forgiving stand-in for a human one. Selectively drugging a 90-amino-acid protein that lives inside cells is hard, and no such molecule exists yet. A single mechanistic paper on a newly named micropeptide is a starting hypothesis, not a validated pathway. Lung cancer resistance runs through many routes, and this is one more added to the map, not the master key.
What makes it worth reading is the shape of the finding. The field has spent years assuming that the interesting resistance biology lived in mutations of the drug's target or in known bypass receptors. Here the culprit is a protein the genome was not supposed to be making at all, translated from a loop of RNA that was supposed to be inert. If circTLL1-90aa holds up in human tumors, the lesson is less about lung cancer than about how much working protein is hiding in the parts of the genome that drug hunters have been skipping over.