In 2018 researchers spotted a tiny RNA fragment called T3p that appeared only in breast‑cancer tissue. That odd observation sparked a multi‑year quest to hunt down similar orphan non‑coding RNAs—now termed oncRNAs—across many tumor types, evaluate which ones actually push disease forward, and test whether they can be caught in a simple blood draw.
The new paper walks through every step of that journey: starting with massive cancer‑genome datasets, building machine‑learning classifiers, running high‑throughput mouse experiments, and finally confirming that oncRNA signals appear in the serum of almost 200 breast‑cancer patients.
OncRNAs Pop Up in Every Major Cancer
Scanning small‑RNA sequencing data from The Cancer Genome Atlas, the team identified roughly 260 000 RNA molecules that were exclusive to tumors. These oncRNAs showed up in all 32 cancer types examined, yet each cancer displayed a distinct expression fingerprint.
For instance, lung tumors carried a different set of oncRNAs than breast tumors. By feeding these patterns into a classifier, the algorithm correctly labelled the cancer type in 90.9 % of cases, and still achieved 82.1 % accuracy on an independent set of 938 tumors.
Even within a single disease, the RNA signatures varied. Basal‑like breast cancers bore a different oncRNA repertoire from luminal‑like counterparts, hinting at yet‑unrecognized subgroups. In short, oncRNAs act like digital barcodes that capture tumor identity at the level of organ, subtype, and cellular state.
Some OncRNAs Drive Tumor Growth
Beyond their diagnostic promise, the researchers asked whether a subset of oncRNAs could actually fuel cancer. They constructed libraries of about 400 candidates drawn from breast, colon, lung and prostate cancers, then introduced each into cultured tumor cells using lentiviral vectors. Half the libraries boosted oncRNA levels; the other half silenced them with “Tough Decoy” constructs.
Modified cells were implanted into mice to see which RNAs accelerated tumor formation. Roughly five percent produced a noticeable effect. Two breast‑cancer oncRNAs stood out: one sparked epithelial‑mesenchymal transition, a key step toward metastasis, while the other ramped up E2F‑target genes, pushing cells into rapid division. Both dramatically hastened tumor growth and spread in separate mouse models.
When the scientists examined patient‑derived TCGA data, tumors that naturally expressed these RNAs displayed the same pathway signatures, reinforcing the link between oncRNA activity and aggressive disease.
OncRNAs Flood the Bloodstream
Perhaps the most clinically relevant finding was that many oncRNAs leak out of cancer cells and travel in the circulation. An analysis of cell‑free RNA from 25 cancer cell lines (spanning nine tissue origins) showed that about a third of oncRNAs are actively secreted.
To test real‑world relevance, serum from 192 breast‑cancer participants in the I‑SPY 2 neoadjuvant trial was profiled before and after chemotherapy. The researchers summed all detected oncRNAs into a single “oncRNA burden” metric. Patients whose post‑treatment serum retained a high burden faced a nearly four‑fold higher risk of death, even after adjusting for standard clinical factors such as pathologic complete response.
Detecting such a powerful signal from just one millilitre of serum was unexpected and suggests that RNA‑based liquid biopsies could outperform DNA‑based tests, especially when tumors shed scant DNA in early disease.
Implications for Minimal Residual Disease
Current approaches to tracking minimal residual disease (MRD) in breast cancer rely heavily on circulating tumor DNA, which can be scarce. Because cancer cells appear to package and release RNA purposefully, oncRNA measurements may provide a more sensitive window into lingering disease after surgery or therapy.
Future Directions
Key questions remain: How exactly do functional oncRNAs exert their influence—through protein partners, RNA‑RNA interactions, or other mechanisms? Can real‑time oncRNA monitoring guide treatment switches, flag recurrence earlier, or improve patient stratification?
Industry partners are already moving forward. The team is collaborating with Exai Bio to translate the oncRNA signatures into diagnostic kits, leveraging AI‑driven models and diverse datasets to sharpen cancer detection.
Behind every data point lies a person who volunteered blood and tissue for research. Respecting that contribution fuels the drive to make oncRNA tools widely available, accelerate discovery, and ultimately improve outcomes for cancer patients worldwide.