This mechanism differs to previously described enhancer hijack events, which bring enhancers and promoters together through structural rearrangement. And differs to examples of focal deletions that impinge on TAD boundaries as we believe this all happens intra-TAD.

By using CRISPR/Cas9 we disrupted the CRNDE super-enhancer in FTO intron 8 CTCF site deleted cells (ALLSIL). This led to a significant reduction in IRX3 expression, meaning this super-enhancer is indeed hijacked by IRX3.

So what’s the deal with CRNDE? It encodes for a lncRNA that is actively transcribed throughout T cell development and harbours a super-enhancer. HiChIP shows this super-enhancer loops to IRX3 suggesting its hijacked in FTO intron 8 CTCF site deleted T-ALL cells (ALLSIL)

We also baited the IRX3 promoter in our CRISPR/Cas9 edited cells (PF382) with and without the FTO intron 8 CTCF site, observing the same effect of increased contacts with the CRNDE region along with increased transcriptional output of IRX3.

With UMI-4C we baited the CTCF site in FTO wild-type cells (PF382). This showed that IRX3 is ‘tethered’ to FTO intron 8. In the same cells we baited the IRX3 promoter which identified minimal contacts with CRNDE with no transcriptional output of IRX3.

CRISPR/Cas9-mediated deletion of the FTO intron 8 CTCF site in IRX3 negative (PF382) T-ALL cells transcriptionally activated IRX3 in single cell sorted clones and polyclonal populations, with no activation if you only delete the MYB site. This suggested the CTCF site was critical.

By using ddPCR probes against the CTCF and MYB sites we identified CN deletions in a site-specific manner. This showed FTO intron 8 deletions occur in 1.5% of adults and 6.2% of paediatric patients with T-ALL. But have a look at the site-specific calls👀

Amazingly these deletions defined a minimal region of ~155kb. ChIP-seq data from a T-ALL cell line identified CTCF binding and a MYB bound enhancer within this region. Both CTCF and MYB have previously been implicated in mechanisms of aberrant gene expression in T-ALL

Expression data was available for 6/13 T-ALL samples showing these were IRX3+ by RNAseq. Notably, 1 patient exhibited allele specific expression supporting our working hypothesis that IRX3 expression may be driven by a cis-acting genetic lesion

By studying published genomic datasets from primary T-ALL patient samples and T-ALL cell lines we identified 13 T-ALL genomes (12 from patients and 1 cell line) with heterozygous copy number losses impinging on FTO intron 8 – the 3’ contact point to IRX3

IRX3 sits in a single TAD with neighbouring genes FTO, CRNDE and IRX5. HiChIP data from an IRX3 positive cell line identified contacts 3’ to IRX3 within FTO intron 8, and 5’ to IRX3 within the CRNDE/IRX5 locus. We thought these two contact points may hold the clue…

IRX3 is not expressed in normal T-cells. But high IRX3 expression is a feature of T-ALL and observed across differing molecular subtypes of T-ALL. Special mention to Somervaille Lab @cruk-mi.bsky.social who have previously shown it can cause T-ALL in vivo. The question is what’s driving this?

For my new followers! Focal deletions of noncoding regions in cancer genomes can have unexpected consequences. Out now in @bloodjournal.bsky.social, we’ve discovered a novel mechanism of oncogene activation whereby focal deletion of a ‘promoter tether’ leads to aberrant expression of IRX3 in T-ALL.👇