Crispri Essentiality Screen Reveals the Essential Enhancers and Oncogenes Comprising the “Regulome” of the Notch-Collaborating Transcription Factor ETS1 in Human T-ALL

T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive lymphoid malignancy derived from immature T cells. No targeted therapies are currently available. Although over 60% of T-ALL samples carry activating mutations in the Notch signaling pathway, this finding has been difficult to translate clinically as pan-Notch inhibitors have significant toxicities. We previously showed that ETS1 is an important Notch co-binding transcription factor in T-ALL and could represent a safer therapeutic target. ETS motifs are the top-ranked motif next to Notch-RBPJ sites in T-ALL, suggesting a strong collaborative interaction. ETS1 knockdown or deletion impairs T-ALL growth in T-ALL cell lines, patient-derived xenografts, and genetically engineered mouse models, illustrating its importance in T-ALL cell growth. In contrast to Notch inhibition, ubiquitous deletion of Ets1 in murine models does not appear to cause significant toxicity.

To further define the downstream enhancers and oncogenes comprising the “regulome” of ETS1 in T-ALL, we used chromatin profiling data to perform a CRISPRi essentiality screen directed at high-probability ETS1-dependent regulatory elements in the human T-ALL cell line THP-6. The “Activity” model (Fulco et al.) was used to predict highly active regulatory regions. Further criteria included “dynamic” ETS1 binding sites (decreased H3K27Ac and ETS1 signals upon ETS1 knockdown at FDR<0.05) and high H3K27Ac ChIP-seq signal in primary T-ALL samples. This library was transduced into THP-6 cells that express doxycycline inducible dCas9-KRAB. Genomic DNA was sequenced before and after expansion in doxycycline. MAGeCK algorithm was used to score elements for negative selection relative to non-targeting controls. Our screen identified 9 regulatory elements with FDR < 6.0E05 that were highly negatively selected and within -LFC range of pan-essential genes. Notably, we identified a previously uncharacterized ETS1-dependent MYB enhancer as the #1 most negatively selected ETS1-dependent T-ALL regulatory element. Characterization and functional analysis of this enhancer is being reported in a separate abstract at this meeting. 2 other highly negatively selected elements were well-characterized T-ALL enhancers. These included the Notch-MYC enhancer (“N-Me”), which we previously described as ETS1-dependent, and a T-ALL enhancer linked to BCL11B. These data validate the utility of our screen.

We additionally identified three novel regulatory elements that have not been previously linked to T-ALL pathogenesis. One notable element is an intronic enhancer linked by H3K27Ac Hi-ChIP to ETS1, FLI1, and an uncharacterized lncRNA. Repression of this +17kB ETS1 intronic enhancer with targeted CRISPRi suppressed expression of ETS1 by 2-fold and inhibited growth by 3-to-4-fold in THP-6 cells. Thus, this enhancer may participate in positive auto-regulation. We next identified a long-range enhancer linked to the promoter of MSI2 by H3K27Ac Hi-ChIP. MSI2 is a recently described T-ALL oncogene that post-translationally promotes MYC expression. Repression of this enhancer inhibited growth by 2-to-3-fold. Finally, the promoter of DOCK2 was amongst the highest negatively selected elements. DOCK2 is a guanine nucleotide exchange factor (GEF) that is expressed primarily in hematopoietic lineages and promotes actin polymerization. Repression of the DOCK2 promoter suppressed expression of DOCK2 by 10-to-100-fold and inhibited growth by 2-to-5-fold in multiple cell lines. Repressing the DOCK2 promoter had no effect on the nearby pan-essential gene SPDL1. To orthogonally confirm these findings, we created a THP-6-DOCK2-FKBP^F36V degron cell line in which the C-terminus of DOCK2 is tagged with FKBP^F36V. Western blot confirmed knockdown of DOCK2 upon treatment of these cells with dTag-V1. DOCK2 knockdown reduced total filamentous actin by 2-fold and slowed proliferation by 2-to-4-fold when compared with wild-type control. These findings suggest for the first time that cytoskeletal remodeling might drive cell intrinsic ALL proliferation.

Taken together, our studies applied a novel approach to define the downstream “regulome” of ETS1 in T-ALL. We revealed a previously uncharacterized network of novel cell-specific non-coding regulatory elements and genes that are important for T-ALL viability.