Mapping the Glucocorticoid Gene Regulatory Network and Alterations That Contribute to Steroid Resistance in Childhood Acute Lymphoblastic Leukemia

Acute lymphoblastic leukemia (ALL) is the most prevalent childhood cancer and despite improved survival rates, relapsed ALL is still among the most common causes of cancer death in children. Although changes in the expression of specific genes have been linked to chemotherapeutic resistance, relatively little is understood of the pharmacogenomic impact of the noncoding, cis-regulatory landscape governing gene regulation. Glucocorticoids (GCs; i.e. steroids) are a mainstay of contemporary, multi-drug chemotherapy in ALL, and GC resistance is predictive of both relapse and poor clinical outcome in ALL. Because GCs function through activation of glucocorticoid receptor (GR), a nuclear receptor transcription factor that interacts directly with cis-regulatory elements, unveiling the glucocorticoid gene regulatory network (GC-GRN) in leukemia cells is crucial to understanding not only the biological mechanism of apoptosis, but also illuminating gene regulatory mechanisms contributing to GC resistance. To test the hypothesis that alterations to the GC-GRN are important contributors to steroid resistance in ALL, we comprehensively mapped cellular responses to GCs in human ALL cell lines using >100 independent functional genomic datasets. This comprehensive approach uncovered thousands of genes and cis-regulatory elements that were responsive to GCs, and further identified >38,000 high-confidence glucocorticoid response elements (GREs) in the ALL genome. A closer examination of these data revealed GR binding profiles that were consistent with the long-range flexible billboard model of gene regulation. By further integrating our results with genetic and epigenetic data in primary ALL cells from patients enrolled on St. Jude clinical trials, we identified 45 DNA sequence variants associated with ex vivo GC resistance that map to GREs and functionally validated an associated variant within the TLE1 gene locus. We also uncovered 1929 accessible chromatin sites (FDR<0.1) in primary ALL cells that were associated with ex vivo GC resistance, and these GC-resistance accessible chromatin sites were highly enriched at GREs determined from ALL cell lines (p<2.2x10 ^-16). High-throughput pharmacogenomic CRISPRi screening in human ALL cell lines with a library of >10,000 sgRNAs targeting >1000 GR binding events at putative GC-resistance accessible chromatin sites identified a subset of GR binding sites implicated in GC resistance. Overall, these data indicate that GCs initiate pervasive, genome-wide effects on the leukemia epigenome and transcriptome, and that genetic and epigenetic alterations to GREs are mechanisms contributing to GC resistance in childhood ALL.