Pre-Clinical Efficacy of Co-Targeting GFI1/KDM1A and BRD4 or JAK1/2 Against AML and Post-MPN Secondary AML Blast Progenitor Cells

Transcriptional regulators (TFs) involved in cell-growth, differentiation and survival of AML stem/progenitor cells (LSCs) include RUNX1, PU.1, CEBPα, c-Myb and c-Myc. LSD1 (KDM1A) is an FAD-dependent amine-oxidase that demethylates mono and dimethyl histone H3 lysine 4 (H3K4Me1 and H3K4Me2). LSD1 is part of the repressor complexes involving GFI1, CoREST and HDAC1/2, that regulate active super-enhancers/enhancers (SEs/Es) and their target genes, mediating transcriptional repression and differentiation block in LSCs. GFI1 is a zinc-finger transcriptional repressor involved in AML development and differentiation. GFI1 contains an N-terminal domain through which it binds to the CoREST/LSD1/HDAC1/2 complex to regulate differentiation in LSCs. CRISPR-suppressor scanning revealed that enzymatic activity of LSD1 was not required for LSC differentiation, instead disruption of binding of LSD1 to GFI1 and CoREST induced differentiation in LSCs. LSD1 and GFI1 expression correlates with worse prognosis in MDS/AML. In present studies, we demonstrate first-time ever that knockout (KO) or degradation of LSD1 utilizing CRISPR-Cas9 or LSD1-FKBP12^(F36V) and dTAG-13, respectively, disrupted LSD1-binding to GFI1/1B and CoREST, inhibiting colony growth and inducing differentiation markers (CD86 and CD11b) and morphologic differentiation of AML and post-MPN sAML blast progenitor cells (BPCs). CRISPR-mediated knockout of LSD1 in the AML OCI-AML5 and sAML SET2 cells significantly increased the permissive H3K4Me2/3-marked chromatin, reduced H3K27Ac occupancy at SEs/Es (by ChIP-Seq), especially of c-Myc and CDK6, as well as repressed DNMT1, CoREST, c-Myc, CDK6, and c-KIT, while inducing GFI1, PU.1, CEBPα, p21, CD11b, and CD86 levels (log2 -fold change by RNA-Seq and by Western analyses). This correlated with growth inhibition, % differentiation and apoptosis of AML and sAML cells. CRISPR-mediated GFI1-KO ± the irreversible LSD1 inhibitor (LSD1i) (INCB059872, INCB), repressed GFI1 levels, yet enhanced expressions of PU.1, p21 and CD11b and significantly increased % morphologic differentiation. Treatment with INCB (0.25 to 1.0 µM) also disrupted binding of LSD1 to GFI1 and to CoREST, increased GFI1/1B and PU.1 and repressed c-Myc protein levels, while significantly inhibiting colony growth, inducing differentiation and loss of viability of AML and post-MPN sAML (SET2 and HEL92.1.7) cells, as well as patient-derived AML and post-MPN sAML blasts (p < 0.01). Following INCB treatment, ATAC-Seq analysis demonstrated gained peaks in GFI1 and PU.1-target genes. Following H3K27Ac ChIP-seq analysis rank-ordering of SEs (ROSE) plot highlighted active SEs of RUNX1, GFI1, BCL2, PU.1, IRF8 and SMYD3, accompanied by increased H3K27Ac occupancy at the chromatin of GFI1 and PU.1 targets. Notably, INCB treatment also increased BRD4 occupancy, especially at the GFI1 and PU.1-target genes. RNA-Seq analysis showed that INCB treatment perturbed mRNA expressions, with positive normalized enrichment scores (NES) for interferon α, inflammatory-response, GFI1-targets and E2F-target gene-sets, and negative NES for c-Myc-targets and oxidative-phosphorylation gene-sets. RNA-Seq analyses of INCB-treated compared to untreated OCI-AML5 and SET-2 cells also demonstrated log2 fold-increase in the mRNA expressions of GFI1, PU.1 and CEBPα target-genes. Utilizing a protein domain-scanning CRISPR-Cas9 sgRNA screen followed by LSD1i treatment, present studies also demonstrate co-dependencies, including BRD4, in AML cells. BET inhibitor (BETi) treatment also depleted LSD1 protein levels, and co-treatment with the BETi OTX015 and INCB induced synergistic lethality in AML and post-MPN sAML blasts (Combination Indices < 1.0). Pre-treatment with INCB re-sensitized JAKi-resistant sAML cells to ruxolitinib-induced apoptosis and BETi-resistant post-MPN sAML cells to BETi-induced apoptosis. Notably, co-treatment with INCB (1.5 mg/kg) and ruxolitinib (20 mg/kg) or OTX015 (50 mg/kg), administered orally for 21 days, compared to ruxolitinib alone or vehicle control, significantly reduced the sAML burden and improved survival of immune-depleted mice engrafted with luciferized sAML HEL92.1.7 xenografts (p < 0.01). Collectively, these findings support further pre-clinical development of LSD1i-based combinations with ruxolitinib and BETi against post-MPN sAML.