Dynamic induction of drug resistance through a stress-responsive enhancer in acute myeloid leukemia [CHIP-seq]

The drug efflux pump ABCB1 is a key driver of chemoresistance, and high expression predicts for treatment failure in acute myeloid leukemia (AML). We show that both acute and chronic exposure of leukemia cells to daunorubicin activates an integrated stress response-like transcriptional program to induce ABCB1 through remodeling and dynamic activation of an ATF4-bound, stress-responsive enhancer. In primary human AML, stress-responsive ABCB1 enhancers are accessible and acetylated, and exposure of fresh blast cells to daunorubicin induces ABCB1 in a dose-dependent manner. Dynamic induction of ABCB1 by diverse stressors, including chemotherapy, facilitates escape of leukemia cells from targeted third-generation ABCB1 inhibition. Stress-induced up regulation of ABCB1 is mitigated by combined use of pharmacologic inhibitors U0126 and ISRIB, which inhibit stress signaling. Overall design: TS75 - K562 cells were purchased from DSMZ (Braunschweig, Germany) and cultured in RPMI 1640 medium (Sigma Aldrich) supplemented with 2mM L-Glutamine (Life Technologies) and 10% fetal bovine serum (Sigma Aldrich). Three separate vials of early passage K562 cells were thawed and cultured separately for two weeks. (Figure 1A). Sensitive cell lines were designated K562_S1-3, aliquots were frozen and stored before commencing selection with 10nM daunorubicin (Sigma Aldrich). Whilst under selection, cells were counted and replated twice per week. The daunorubicin concentration was adjusted depending on whether cell numbers were increasing (dose increased by 25-50%), decreasing (dose reduced by 25-50%) or static (dose unchanged). Selection continued until cell lines were capable of expansion in medium containing 500nM daunorubicin. Resistant cell lines were designated K562_R1-3. The time taken to acquire this level of resistance was 106 days (K562_R1 and R3) and 117 days (K562_R2). K562_R1-3 were culture without daunorubicin for 10 days prior to RNA extraction, assessment of daunorubicin IC50 and freezing of aliquots. All cell lines (K562_S1-3, K562_R1-3) were tested for mycoplasma and authenticated by short tandem repeat DNA profiling. H3K27ac ChIP sequencing was performed as detailed below. Data from TS75 was used to identify putative ABCB1 enhancers by comparing resistant (K562_R1-3) with sensitive (K562_S1-3) cells. TS93 - To confirm that putative enhancers were functional, we next performed targeted silencing using a CRISPR-dCas9-KRAB system. We designed multiple sgRNAs for each region and screened them in K562_R3 cells. The most active guides were then selected for use in all resistant cell lines. K562_R1-3 cells were dual infected with pHR-SFFV-dCas9-BFP-KRAB and pLKO5.sgRNA.EFS.tRFP657, the latter expressing an sgRNA targeting an enhancer (E1-4) or the promoter, or a non-targeting control (EV). We used ChIPseq for H3K9 trimethylation to confirm that silencing was discrete and accurate, this was performed as described below. TS98 - To examine enhancer usage in primary AML we performed H3K27ac ChIP sequencing on cryopreserved bulk AML samples, each with their own unique identifier (BBxxx).