Parkus, S. | Search bioRxiv

bioRxiv2026

Aberrant chromatin looping by NUP98-HOXA9 is constrained by CTCF and facilitated by cohesin

The acute myeloid leukemia (AML) fusion protein NUP98-HOXA9 (NHA9) drives leukemogenesis by promoting aberrant chromatin loop formation through phase separation, yet the mechanisms underlying these interactions remain unclear. To address this, we dissect the interplay between NHA9 and individual loop extrusion factors using in situ Hi-C, CUT&RUN, RNA-seq, and Auxin-inducible degradation of CTCF or RAD21. CTCF was found to be dispensable for NHA9 loop formation, although CTCF binding constrained a subset of loops that emerged only upon CTCF depletion. In contrast, cohesin played a distance-dependent role where short-range NHA9 loops formed independently of RAD21, while long-range loops were strongly cohesin-dependent. Despite this requirement, RAD21 showed minimal enrichment at NHA9 loop anchors, indicating that NHA9 does not function as a canonical cohesin barrier. Instead, these findings support a non-canonical model in which cohesin transiently facilitates interactions between distal NHA9-bound loci, which are subsequently stabilized through NHA9 phase separation. Together, this work reveals a distinct mechanism of oncogenic chromatin looping in which NUP98-HOXA9 cooperates with canonical loop extrusion machinery to reprogram genome architecture in AML.

Molecular Biology↗

bioRxiv2024

Response splicing QTLs in primary human chondrocytes identifies putative osteoarthritis risk genes

Osteoarthritis affects millions worldwide, yet effective treatments remain elusive due to poorly understood molecular mechanisms. While genome-wide association studies (GWAS) have identified over 100 OA-associated loci, identifying the genes impacted at each locus remains challenging. Several studies have mapped expression quantitative trait loci (eQTL) in chondrocytes and colocalized them with OA GWAS variants to identify putative OA risk genes; however, the degree to which genetic variants influence OA risk via alternative splicing has not been explored. We investigated the role of alternative splicing in OA pathogenesis using RNA-seq data from 101 human chondrocyte samples treated with PBS (control) or fibronectin fragment (FN-f), an OA trigger. We identified 590 differentially spliced genes between conditions, with FN-f inducing splicing events similar to those in primary OA tissue. We used CRISPR/Cas9 to mimic an SNRNP70 splicing event observed in OA and FN-f-treated chondrocytes and found that it induced an OA-like expression pattern. Integration with genotyping data revealed 7,188 splicing quantitative trait loci (sQTL) affecting 3,056 genes. While many sQTLs were shared, we identified 738 and 343 condition-specific sQTLs for control and FN-f, respectively. We identified 15 RNA binding proteins whose binding sites were enriched at sQTL splice junctions and found that expression of those RNA binding proteins correlated with exon inclusion. Colocalization with OA GWAS identified 6 putative risk genes, including a novel candidate, PBRM1. Our study highlights the significant impact of alternative splicing in OA and provides potential therapeutic targets for future research.

genetics↗