Back

Long Noncoding RNAs Preserve Pancreatic Cancer Identity and Resist Cell Fate Conversion

Grygoryev, D.; Lee, S.-W.; Frankston, C. M.; Rakshe, S.; Berry, M.; Hirano, A.; Ekstrom, T.; Manalo, E.; Tessier, J.; Chow-Castro, M.; Link, J.; Keith, D.; Sheppard, B. C.; Fei, S.; Morgan, T.; Remotti, H. E.; Yang, W.; Furth, E.; Thakurela, S.; Sears, R. C.; Kim, J.

2025-08-13 cancer biology
10.1101/2025.06.25.661419 bioRxiv
Show abstract

The Yamanaka factors (OCT4, SOX2, KLF4, and MYC; OSKM) can rejuvenate aging phenotypes in somatic cell types by resetting the epigenetic landscape. Curiously, most solid tumor cells remain largely resistant to reprogramming despite their well-documented plasticity, and the underlying mechanisms are unclear. Here, we combined genomic profiling and in vivo assays to investigate OSKM-mediated reprogramming of pancreatic ductal adenocarcinoma (PDAC). In the initial stages, we found that cancer-specific genes were refractory while mesodermal/ECM programs, normally silenced by PRC2, were aberrantly upregulated. A CRISPR interference screen for OSKM reprogramming coupled with functional analyses revealed that suppression of cancer-associated long noncoding RNAs (lncRNAs) erased malignant epithelial programs, restored tumor suppressor activity, and impaired tumorigenicity in vivo. We further identified that ATXN7L3-AS1 lncRNA sustains the PDAC malignant identity through its association with active epithelial oncogenic programs and poised PRC2-targeted developmental genes, thereby supporting both plasticity and memory. Thus, by exploring why cancer cells are resistant to reprogramming, we identify lncRNAs as gatekeepers of malignant identity, suggesting that targeting lncRNAs could be a generalizable therapeutic strategy in treating solid tumors.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.