Back

KDM5-driven transcriptional noise fuels plasticity-led awakening and relapse in paediatric cancer

Allo Anido, A.; Roux, C.; Chen, E.; Hamer, S.; Shea, A.; Sadr, A. S.; English, C.; Butterworth, C.; Che, H.; Bellini, A.; Geoerger, B.; Schleiermacher, G.; Chesler, L.; Morgan, M. D.; Bruna, A.

2025-10-14 cancer biology
10.1101/2025.10.13.682004 bioRxiv
Show abstract

How drug-tolerant persister (DTP) cells escape quiescence to drive tumour relapse is a central unresolved question in cancer evolution. Here, we identify transcriptional noise (TN), defined as the stochastic variability in gene expression, as a latent property of paediatric cancer cells that becomes a driver of adaptive regrowth after treatment withdrawal. Using functional assays, lineage tracing, single-cell transcriptomics, and multiscale landscape modelling, we show that therapy enriches mesenchymal-like tolerant states in neuroblastoma without clonal selection, while post-treatment awakening is a stochastic process fuelled by noise-enabled plasticity in cell-identity programmes. The histone demethylase KDM5A relocates to noisy cell-state genes during awakening, promoting H3K4me3 removal and chromatin remodelling at these loci. KDM5 inhibition abrogates this process, and suppresses transcriptional noise, halts DTP exit, and prevents tumour recovery in both neuroblastoma and hepatoblastoma models. These results establish DTP as an exploitable evolutionary bottleneck, positioning KDM5-mediated transcriptional noise as an actionable therapeutic target to limit cancer adaptation and relapse GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=120 SRC="FIGDIR/small/682004v1_ufig1.gif" ALT="Figure 1"> View larger version (34K): org.highwire.dtl.DTLVardef@edba70org.highwire.dtl.DTLVardef@16b1bd9org.highwire.dtl.DTLVardef@f153acorg.highwire.dtl.DTLVardef@1ac5a4e_HPS_FORMAT_FIGEXP M_FIG C_FIG

Matching journals

The top 5 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.