Paper Details

ABSTRACT

Background 

Children of Hispanic/Latino ethnicity have a 1.4-fold and 2-fold higher risk of developing ALL than their Non-Latino White and Black counterparts, respectively (Feng et al., Am J Epidemiol, 2021). These differences in incidence are unlikely to be explained by acquired mutations alone, highlighting the need to identify ancestry-associated genetic determinants of susceptibility. Genome-wide association studies identified noncoding genetic variants at the ARID5B locus, including rs7090445, as reproducibly associated with increased childhood B-cell acute lymphoblastic leukemia (B-ALL) incidence, but the underlying mechanism remains undefined. ARID5B is a chromatin regulator that binds AT-rich DNA and participates in transcriptional and epigenetic control through interactions with chromatin-modifying proteins (Baba et al., Nat Cell Biol, 2011; Kutschat et al., Nucleic Acids Res., 2026). Recent studies identified rs7090445 as an intronic enhancer variant at the ARID5B locus, and the risk allele is associated with reduced ARID5B expression (Zhao et al., J Natl Cancer Inst., 2022; Studd et al., Nat Commun, 2017); however, the transcriptional regulators altered by the rs7090445 risk allele remain unclear.

Methods and Results

To further validate the association between rs7090445 variants and ARID5B mRNA levels in primary human B-ALL, we analyzed ~300 primary B-ALL RNA-seq samples with known rs7090445 variant genotypes. Moreover, we further characterized B-ALL samples by their predicted cell-type of origin using published gene expression profiles of human B cell progenitor development (Iacobucci et al., Nature Cancer, 2025). Here, the presence of the rs7090445 risk allele is associated with reduced ARID5B mRNA expression in pro-B and pre-B ALL, suggesting a potential stage-specific role of rs7090445 in transformed B-ALL.

To understand the precise molecular mechanisms underpinning ARID5B mRNA reduction by the rs7090445 risk allele, we used a CRISPR prime editing to generate clonally derived isogenic Nalm6 and Reh B-ALL cell lines carrying homozygous non-risk (T/T), heterozygous risk (T/C), or homozygous risk (C/C) rs7090445 genotypes. In isogenic B-ALL models, the rs7090445 risk allele was sufficient to reduce ARID5B mRNA levels in both heterozygous and homozygous risk-allele-bearing cells relative to homozygous non-risk clones, as evidenced by decreased H3K36me3 across the ARID5B gene body in risk allele-bearing clones, consistent with impaired transcriptional elongation. In addition, ATAC-sequencing and Capture Hi-C analyses found reduced chromatin accessibility and enhancer-promoter contacts at the intronic enhancer harboring the rs7090445 risk allele. These data suggest the rs7090445 risk allele reduces intronic enhancer activity.

Previous studies suggest that the rs7090445 risk allele reduces the binding of transcriptional activators, such as MEF2C, to the ARID5B intronic enhancer (Zhao et al., J Natl Cancer Inst., 2022). Using transcriptional motif analysis and ENCODE ChIP-seq data at the rs7090445 locus, we found that the rs7090445 risk allele introduced a CUX1 binding motif absent in non-risk allele-bearing genotypes. To functionally validate this observation, we isolated nuclear extracts from Nalm6 and Reh cells and performed a biotinylated DNA oligo pull-down of the intronic enhancer sequence harboring the risk and non-risk alleles. Here, we found enhanced MEF2C binding upon non-risk allele-bearing oligo pulldown, consistent with previous studies. Notably, risk allele-bearing oligo pulldown showed increased CUX1 binding, supporting enhanced CUX1 binding in risk allele-bearing cells. Functionally, CUX1 knockout in homozygous risk cells restored ARID5B mRNA expression to non-risk levels and increased enhancer accessibility, supporting a repressive role for CUX1 at this locus.

Conclusions