Dissecting mechanisms at an IKZF1 cis-regulatory element harboring ancestry-specific and cross-ancestry B-cell acute lymphoblastic leukemia risk variation

Abstract

Abstract B-cell acute lymphoblastic leukemia (B-ALL) is the most common pediatric cancer, with Hispanic/Latino (H/L) children in the United States exhibiting the highest incidence rates. Despite this, the genetic basis of this susceptibility remains poorly understood. Genome-wide association studies identified several well-replicated single-nucleotide polymorphisms (SNPs) associated with childhood B-ALL risk. Recent analyses discovered a novel childhood B-ALL risk signal near the IKZF1 gene that is positively selected for in H/L populations. The putatively causal variant, rs1451367 C>T, resides downstream of IKZF1 in a 3'cis-regulatory element (3'CRE) defined by histone marks and chromatin accessibility. In addition, an independent B-ALL risk variant, rs17133807 G>A, found across ancestries lies 26 bp away in the same regulatory element. However, how this regulatory element mechanistically impacts IKZF1 expression remains unknown. To unravel the molecular role of the 3'CRE, we integrated chromatin accessibility data, 3D chromatin interaction mapping, computational transcription factor (TF)-binding predictions, and functional genomic assays. ATAC-seq profiling from 156 B-ALL patients revealed a distinct and variable accessible chromatin landscape at this 3'CRE in B cell progenitors and precursors, suggestive of highly-regulated and stage-specific activity. Of note, in the presence of the risk variants, allele-specific chromatin accessibility was reduced, implicating decreased enhancer activity. Further, Promoter Capture Hi-C data from human lymphoblastoid cells revealed strong interaction with the IKZF1 promoter, supporting the CRE's role as a bona fide IKZF1 enhancer. Having established its accessibility and promoter interaction in early B cells, we next sought to identify the TFs occupying the 3'CRE and determine how their binding is altered by the risk allele. Employing motifbreakR and delta SVM modeling, we predicted that rs1451367 would disrupt binding of several Erythroblast Transformation Specific family members, including ERG, ETV6, ELK3, in addition to altering a potential IKZF1 motif. Prioritizing these TFs with a negative differential binding score, we then performed chromatin immunoprecipitation (ChIP) coupled with quantitative real-time PCR (qPCR) to examine differential TF binding. To model the risk variants at this locus, we used CRISPR/Cas9 genome editing to introduce a 60 bp microdeletion spanning the 3'CRE variants in SEM cells, a human B-ALL cell line with a KMT2A-AFF1 translocation, in which we confirmed the CRE harbors accessible chromatin. ChIP qPCR revealed strong ERG occupancy at the 3'CRE, which was abolished in the presence of microdeletions of the risk variants. These findings implicate ERG as a key participant in orchestrating IKZF1 expression and demonstrate that the risk variant disrupts this interaction. Intriguingly, Nalm-6 cells, which harbor an IGH-DUX4 rearrangement, displayed a strikingly different pattern, with an absence of ERG binding at the 3'CRE. Western blotting detected an additional ERG isoform in these cells, which we identified as ERGalt – a dominant-negative isoform generated through alternative transcription driven by IGH-DUX4. ChIP across multiple B-ALL lines confirmed distinct binding profiles, underscoring the context-dependent regulation of this enhancer by ERG across leukemic subtypes. Notably, as IKZF1 itself is known to regulate ERG, our findings suggest a previously unrecognized feedback mechanism that might occur among these TFs. Also, we found the adjacent multi-ancestry risk variant to disrupt a TCF3 motif, with ongoing studies assessing its impact on IKZF1 transcription. Thus, to more finely dissect precise regulation at this element, we are conducting systematic mutagenesis studies at this CRE. Collectively, this work identifies a mechanistic model in which the Hispanic-specific risk variant disrupts ERG-mediated regulation of IKZF1, stalling B-cell differentiation at an early stage and priming cells for leukemic transformation. By identifying ERG as a novel transcriptional regulator at the IKZF1 3'CRE, our work reveals a previously unrecognized regulatory intersection between two key B-cell transcription factors. These insights not only elucidate how population-specific noncoding variation contributes to B-ALL susceptibility, but also highlight a regulatory axis that could be leveraged for enhancer-targeted or TF-modulating therapeutic interventions in the future.