The Transcriptional Landscape of Ph+B-ALL Is Orchestrated By Long-Range Enhancer-Promoter Interactions and the Coordinated Action of Phosphorylation-Dependent and Phosphorylation-Independent Transcription Factors

Abstract

Background and Hypothesis: B-cell precursor leukaemia (B-ALL) subtypes can be distinguished by their initiating genetic lesions. Likewise, each subtype exhibits a unique transcriptional program that drives the malignancy and defines the subtype. This often relates to the initiating genetic alteration that affects genes encoding transcription factors (TFs). However, for B-ALLs not driven by TF alterations it is less clear how the transcriptional program is established. Ph+B-ALL is a historically poor prognosis B-ALL driven by the BCR::ABL1 oncogene, a constitutive tyrosine kinase that activates its targets via phosphorylation. We hypothesised that BCR::ABL1 initiates the transcriptional signature of Ph+B-ALL cells by a two-step mechanism involving phosphorylation-activatable TFs, which then induce the expression of additional TFs that consecutively establish the Ph+B-ALL-defining transcriptome. As gene expression is often modulated by cis-regulatory elements such as enhancers, we further speculated that BCR::ABL1 reprograms the enhancer landscape to promote the transcriptional changes that define Ph+B-ALL. Results: Using an in vitro model of BCR::ABL1-induced transformation of murine B-cell precursors, we show that transcriptional deregulation by BCR::ABL1 is a continuous process evolving until full transformation is established. Only a minority of induced transcriptional changes occur immediately upon BCR::ABL1 expression (3 days), while most affected genes require an extended period of BCR::ABL1 to become deregulated. Furthermore, less than half of deregulated genes in murine Ph+B-ALL and less than 10% of human Ph+B-ALL-defining genes were sensitive to 24h BCR::ABL1 inhibition, suggesting an involvement of phosphorylation-dependent and phosphorylation-independent TFs. TF motif enrichment analysis of active chromatin in human and mouse Ph+B-ALL cells indicated the phosphorylation-dependent TF STAT5 and the phosphorylation-independent TF ETV5 as main mediators. Degron-induced rapid degradation of each TF in Ph+B-ALL cells confirmed their unique and overlapping roles and coordinated action. In parallel to gene expression changes, Ph+B-ALL cells further acquire changes in their enhancer signatures defined by H3K27ac+ at non-promoter regions, and these enhancer signatures can be efficiently used to distinguish Ph+B-ALL from other B-ALL subtypes. Analysis of chromatin interactions of these active enhancers with respective active promoters by Promoter-Capture Hi-C (PCHI-C) in combination with H3K27ac-ChIP-Seq showed that most active genes in Ph+B-ALL cells physically interact with an enhancer, including many genes described to be essential for Ph+B-ALL cells, such as CCND2, BCL2, BCL2L1, and ETV5. Using PCHI-C and H3K27ac-HiChIP we further show that, like gene expression and enhancer signatures, physical contacts of promoters and enhancers can be efficiently used to identify Ph+B-ALL cells and distinguish them from other B-ALLs such as KMT2A-rearranged B-ALL. Subtype-specific chromatin interactions that allow this differentiation directly relate to the transcriptional programs that define the respective B-ALL subtype. Conclusion: We describe here the molecular details of how signals from an oncogenic tyrosine kinase become converted into changes in gene expression that define the B-lineage leukaemia subtype Ph+B-ALL. Our work highlights the complexity of these changes by acknowledging the role of gene regulatory elements/enhancers in this process and further describe for the first time that B-ALL subtypes can be distinguished by their enhancer usage and promoter-enhancer interactions. Uncovering the precise enhancer regions for the genes deregulated in Ph+B-ALL and the TFs recruited to them may open new windows for therapeutic intervention.