Preclinical efficacy of targeting epigenetic mechanisms in AML with 3q26 lesions and EVI1 overexpression

Abstract

AML with chromosomal alterations involving 3q26 overexpress the transcription factor (TF) EVI1, associated with therapy refractoriness and inferior overall survival in AML. Consistent with a CRISPR screen highlighting BRD4 dependency, treatment with BET inhibitor (BETi) repressed EVI1, LEF1, c-Myc, c-Myb, CDK4/6, and MCL1, and induced apoptosis of AML cells with 3q26 lesions. Tegavivint (TV, BC-2059), known to disrupt binding of nuclear β-catenin and TCF7L2/LEF1 with TBL1, also inhibited co-localization of EVI1 with TBL1 and dose-dependently induced apoptosis in AML cell lines and patient-derived (PD) AML cells with 3q26.2 lesions. TV treatment repressed EVI1, attenuated enhancer activity at ERG, TCF7L2, GATA2 and MECOM loci, abolished interactions between MYC enhancers, repressing AML stemness while upregulating mRNA gene-sets of interferon/inflammatory response, TGF-β signaling and apoptosis-regulation. Co-treatment with TV and BETi or venetoclax induced synergistic in vitro lethality and reduced AML burden, improving survival of NSG mice harboring xenografts of AML with 3q26.2 lesions.

Introduction:

EVI1 (Human Ecotropic Viral Integration Site 1) is a zinc finger transcriptional regulator encoded from the MDS1 (Myelodysplastic Syndrome 1) and EVI1 complex (MECOM) locus at chromosome 3q26.2 (1–4). Conditional knockout of Evi1 results in a marked reduction in long-term hematopoietic stem cells (LT-HSCs), and failure to engraft and repopulate upon transplantation (5). The self-renewal ability of LT-HSCs is linked to EVI1 expression, and many LT-HSC associated genes have EVI1 binding sites in their regulatory regions (6,7). EVI1 transcripts are decreased in human CD34+ cells after stimulation of differentiation induced by cytokine administration, suggesting that downregulation of EVI1 is an important step in terminal differentiation of many hematopoietic lineages (8,9). For most myeloid lineages, EVI1 functions to promote a stem or early progenitor transcriptional program (8–11). Forced expression of EVI1 caused transcriptional repression of the lineage specific gene myeloperoxidase and the myeloid transcription factors C/EBPα (CCAAT enhancer binding protein alpha) and RUNX1 (RUNX Family Transcription Factor 1) (10,12). EVI1 blocks erythroid differentiation through binding and inhibiting transcriptional activity of the myeloid transcription factor GATA1 (GATA Binding Protein 1) and upregulation of PU.1 (Spi-1 Proto-Oncogene) (13–15). EVI1 is overexpressed in approximately 5% of AML, including those harboring chromosome alterations involving relocation of enhancers (16,17). These include GATA2 enhancer in inv3/t(3;3) (q21;26), MYC enhancer in t(3;8) (q26;q24), or RUNX1 enhancer in t(3:21) (q26;q22) (17–20). In inv3/t(3;3) where GATA2 enhancer (3q21) is repositioned to cause EVI1 overexpression, the tumor suppressor GATA2 is simultaneously repressed (18,19). GATA2 repression, coupled with increased activity of EVI1 in AML with chromosome 3q26 lesions, promotes disease progression and an aggressive phenotype (21,22). This is recognized by the WHO as a separate AML subtype that correlates with poor prognosis and clinical outcome (23,24), exhibiting chemo-resistance and 2-year survival of less than 10% (24–26). Therefore, there is a glaring need to develop and test targeted therapies with improved efficacy against AML with EVI1 overexpression. Among the targets of EVI1 are the ETS family transcription factor ERG, MYC, KIT, Bcl-xL and MPL (27–30). Recently, in addition to being a transcriptional target of EVI1, ERG was shown to contribute to EVI1 mediated myeloid differentiation block, AML stemness and anti-apoptotic effects (27). Other mutations that commonly co-occur with t(3;3) or inv(3) and EVI1 overexpression include mutations in NRAS and SF3B1 (31–33). Additionally, EVI1 overexpression due to chromosome 3q26.2 lesions in MDS and AML is also frequently associated with monosomy 7, which contributes to poorer relapse-free and overall survival (33–35).

A previous report had highlighted that CML blast crisis cells express high β-catenin levels, and knockdown of β-catenin and its related co-transcription factor LEF1 (lymphoid enhancer factor 1) was shown to attenuate EVI1 levels (36,37). Consistent with this, two tandem LEF1-β-catenin binding sites are present upstream of EVI1 which are bound by LEF1 (37). We previously reported that treatment with tegavivint (TV) disrupts the binding of nuclear β-catenin to the scaffold proteins TBL1/R1 (Transducin Beta Like 1 X-Linked/TBL1X Receptor 1) complexed to TCF7L2 (Transcription Factor 7 Like 2, TCF4), a member of LEF1 family of transcription factors, which reduced the nuclear β-catenin levels and inhibited targets of the transcriptional complex, including c-Myc (37,38). This resulted in growth inhibition and apoptosis of AML stem progenitor cells (38). A previous report had demonstrated that, due to increased BRD4 (Bromodomain Containing 4) occupancy on the rearranged GATA2 enhancer driving EVI1 overexpression, treatment with a BET protein inhibitor (BETi) represses EVI1 and induces loss of viability of EVI1 overexpressing AML cells (17). Related to this, our previous studies had also demonstrated that treatment with TV overcomes the non-genetic mechanism of tolerance-resistance to BETi, by inhibiting emergence of c-Myc re-expression in BETi-tolerant/resistant AML cells and restoring their sensitivity to BETi (39). Taken together, and based on the rationale generated by these reports, we determined here effects of TV on active chromatin, the transcriptome and protein expressions in AML cell-models harboring 3q26.2 lesions, where EVI1 overexpression promotes differentiation-arrest, cell growth and survival of AML stem/progenitor cells. We also determined whether EVI1 interacts with TBL1 and treatment with TV disrupts the binding of TBL1 to β-catenin, TCF7L2 and EVI1. Finally, we also evaluated whether co-treatment with TV and the BETi or venetoclax would exert synergistic in vitro cell lethality and induce superior in vivo anti-AML efficacy than treatment with each of the drugs alone against xenograft models of human AML harboring t(3;3) or inv(3) with EVI1 overexpression.

Materials and Methods:

Cell lines and cell culture.

MUTZ-3 [RRID:CVCL_1433], UCSD-AML1 [RRID:CVCL_1853], and HNT-34 [RRID:CVCL_2071] cells were obtained from the DSMZ. OCI-AML-20 [RRID:CVCL_UE33] and OP9 feeder cells [ATCC Cat# CRL-2749, RRID:CVCL_4398] were a kind gift from Dr. Mark Minden (Princess Margaret Cancer Centre, Toronto, Ontario, Canada; Department of Medical Biophysics, University of Toronto, Ontario, Canada). AML191, AML194 and AML219 were generated in our laboratory from primary patient samples and were verified by the MD Anderson Pathology department as inv(3)(q21;q26.2) by FISH, with additional cytogenetics determined by Array-CGH and chromosomal spreads. All experiments with cell lines were performed within six months after thawing or obtaining from DSMZ. The cell lines were also authenticated in the Characterized Cell Line Core Facility at M.D. Anderson Cancer Center, Houston TX. UCSD-AML1, HNT-34, AML191, AML194, and AML219 were cultured in high-glucose-formulated RPMI-1640 with 20% FBS, 1% non-essential amino acids (NEAA), and 1% penicillin/streptomycin. UCSD-AML1 and PD AML cell lines were grown in the presence of 10 ng/mL GM-CSF ([Cat # 300–03] PeproTech, Cranbury, NJ). MUTZ-3 cells were cultured in alpha-MEM with 20% FBS, 1% NEAA, 1% penicillin/streptomycin and 10% 5637 (bladder cancer) cell line [RRID:CVCL_0126]-conditioned media. OCI-AML-20 cells were co-cultured with OP9 cells in IMDM with 20% FBS, 1% NEAA, 1% penicillin/streptomycin and 20 ng/mL GM-CSF (PeproTech). Logarithmically growing, mycoplasma-negative cells were utilized for all experiments. Following drug treatments, cells were washed free of the drug(s) prior to the performance of the studies described.

Cell Line Authentication.

The cell lines utilized in these studies were authenticated in the Characterized Cell Line Core Facility at M.D. Anderson Cancer Center, Houston TX utilizing STR profiling.

RNA isolation and quantitative polymerase chain reaction.

Following the designated treatments, total RNA was isolated from AML cells utilizing a PureLink RNA Mini kit from Ambion, Inc. (Austin, TX) and reverse transcribed with a High Capacity Reverse Transcription kit from Life Technologies (Carlsbad, CA). Quantitative real-time PCR analysis for the expression of target genes was performed on cDNA using TaqMan probes and a TaqMan Universal PCR Mastermix from Applied Biosystems (Foster City, CA). Relative mRNA expression was normalized to the expression of GAPDH and compared to the untreated cells.

Data Availability Statement.

The datasets generated during and/or analyzed during the current study are available in the GEO repository as a Super Series and have been assigned Accession ID: GSE247095.

Additional detailed methods are provided in the Supplemental Materials and Methods.

Results:

BET inhibitor and TV treatment induces loss of viability of AML cells harboring inv(3)/t(3;3).

We first confirmed the in vitro lethal activity of a pan-BET protein inhibitor, OTX015, on AML cell lines (MUTZ-3, UCSD-AML1, OCI-AML20, HNT-34, AML191, AML194 and AML219) as well as in PD AML cells harboring inv(3q26) or t(3:3) and overexpressing EVI1 (1,39,40). Fig. S1A depicts the results of FISH and cytogenetics analyses of AML191, AML194 and AML219 cells, whereas Fig. S1B shows the oncoplot of the genetic alterations in PD AML cells. Fig. 1A demonstrates that treatment with OTX015 dose-dependently induced lethality in the AML cell lines, with AML191 showing less sensitivity than AML-194 cells. HNT-34, UCSD-AML1 and OCI-AML20 cells exhibited intermediate sensitivity. The lethal activity of OTX015 was associated with significant repression of mRNA of MECOM, LEF1, KIT, MYB and MYC but increased expression of HEXIM1 (Fig. 1B and S1C) (39). OTX015 treatment also concomitantly reduced the protein expressions of EVI1, CDK4/6, c-Myb, c-Myc and LEF1, while increasing protein levels of HEXIM1, BRD4, p21, CD11b and cleaved PARP (Fig. 1C, S1D and S1E). Exposure to OTX015 also induced greater loss of viability in PD AML cells harboring inv(3q26) or t(3:3) (mean values, n = 11), as compared to the AML cells lacking it (mean values, n = 8) (p < 0.05) (Fig. 1D and S1F). Since β-catenin acts as the co-factor for LEF1 and reduction in nuclear β-catenin levels would reduce LEF1, and potentially EVI1 levels, we next determined effects of TV, previously documented to attenuate nuclear β-catenin levels (38), on viability and associated gene-expression perturbations in AML cell lines and PD AML cells harboring inv(3q26) or t(3:3). Treatment with TV dose-dependently induced apoptosis in AML cells lines as well as loss of viability in PD AML cells with inv(3q26) or t(3:3) (Fig. 1E and 1F). In contrast, exposure to TV or OTX015 induced significantly less lethality in CD34+ normal progenitor cells (Fig. S1G and S1H). Parenthetically, as previously reported by us, treatment with TV also induced lethality in AML cells that did not harbor inv(3q26) or t(3:3) (Fig. S1I) (38,39). Concomitant with inducing lethality, TV treatment also reduced mRNA expressions of MECOM, LEF1, MYB, MYC and KIT, while inducing mRNA expression of Axin-2 in UCSD-AML1 and PD AML-191 cells (Fig. 1G and S1J) (38). Additionally, treatment with TV reduced protein levels of the active and un-phosphorylated β-catenin (39), LEF1, TCF7L2, EVI1, c-Myb, c-Myc, BCL2 and Bcl-xL, but increased levels of p21, CD11b and cleaved PARP in AML-191 and UCSD-AML1 cells (Fig. 1H and S1K and S1L). Notably, reduction in EVI1 levels caused by TV treatment was not restored by co-treatment with the proteasome inhibitor carfilzomib, suggesting that the effect of TV on EVI1 levels was transcriptional, not mediated through its proteosomal degradation (Fig. S1M). We also determined whether TV-induced lethality was due to increase in reactive oxygen species (ROS) levels (41). Although TV treatment increased intracellular ROS levels concomitant with inducing apoptosis in AML-191 cells, co-treatment with N-acetylcysteine that reduced ROS levels, did not inhibit TV-induced apoptosis in AML-191 cells (Fig. S1N and S1O) (41). We also determined whether exposure to lower concentrations of TV that induce less lethality over 48 hours (Fig. 1E) would induce myeloid differentiation. As shown in Fig. S1P, treatment with low levels of TV for 7 days increased the % of cells with CD11b expression, a myeloid differentiation marker in AML191 cells (11,42). Overall, these data highlight in vitro lethal activity of TV and BETi in AML cells with inv(3q26) or t(3:3).

Figure 1. Treatment with BET inhibitor OTX015 or β-catenin antagonist TV depletes EVI1 and c-Myc expression and dose-dependently induces lethality in 3q26.2-rearranged EVI1-expressing AML cells.

A. 3q26.2-rearranged EVI1-expressing AML cells were treated with the indicated concentrations of OTX015 for 48 hours. At the end of treatment, cells were stained with annexin V and TO-PRO-3 iodide and the % of annexin V-positive, apoptotic cells were determined by flow cytometry. Curves; mean of three or more independent experiments. Error bars; standard error of the mean (S.E.M.). B. UCSD-AML1 cells were treated with 1000 nM of OTX015 for 8 hours. Total RNA was isolated and reverse transcribed. The resulting cDNA was utilized for qPCR and mRNA expression was normalized to GAPDH. C. Representative immunoblot analyses of 3q26.2-rearranged cell lines UCSD-AML1 and PD AML191 treated with the indicated concentrations of OTX015 for 24 hours. The expression of β-Actin in the lysates served as the loading control. D. Patient-derived 3q26.2-rearranged acute myeloid leukemia (AML) samples (n=11) were treated with the indicated concentrations of OTX015 for 48 hours. The percentage of non-viable cells were determined by staining with TO-PRO-3 iodide and flow cytometry. Each dot represents a unique patient sample. E. 3q26.2-rearranged EVI1-expressing AML cells were treated with the indicated concentrations of TV for 48 hours. At the end of treatment, cells were stained with annexin V and TO-PRO-3 iodide and the % of annexin V-positive, apoptotic cells were determined by flow cytometry. Curves; mean of three or more independent experiments. Error bars; S.E.M. F. Patient-derived 3q26.2-rearranged AML samples (n=11) were treated with the indicated concentrations of TV for 48 hours. The percentage of non-viable cells were determined by staining with TO-PRO-3 iodide and flow cytometry. Each dot represents a unique patient sample. G. UCSD-AML1 cells were treated with 200 nM of TV for 16 hours. Total RNA was isolated and reverse transcribed. The resulting cDNA was utilized for qPCR and mRNA expression was normalized to GAPDH. H. Representative immunoblot analyses of 3q26.2-rearranged cell lines treated with the indicated concentrations of TV for 24 hours. The expression of β-Actin in the lysates served as the loading control.

TV disrupts the interaction of LEF1 and EVI1 with TBL1.

As previously reported for AML cells without inv(3q26) (38), here we determined whether treatment of AML cells harboring inv(3q26) with TV would disrupt binding of nuclear β-catenin to the scaffold protein TBL1 and reduce the levels of β-catenin. Utilizing confocal microscopy, we demonstrate that TV treatment reduced co-localization of β-catenin with TBL1 and reduced nuclear β-catenin but not TBL1 levels in AML191 cells (Fig. S2A). The violin plot shows a significant decline in the nuclear β-catenin levels. Notably, via confocal microscopy we also determined that treatment with TV disrupted co-localization of LEF1 with TBL1 and significantly depleted nuclear LEF1 levels in AML191 and in UCSD-AML1 cells, as is also shown in the violin plot (Fig. 2A and S2B). Next, we interrogated whether EVI1 also co-localizes with and binds to TBL1, and whether TV treatment disrupts this interaction. Fig. 2B and 2C demonstrate that indeed TV treatment disrupted co-localization of EVI1 with TBL1 and significantly reduced the levels of nuclear EVI1 in AML191 and UCSD-AML1 cells. To evaluate this important finding further, we also utilized the proximity ligation assay (PLA) (43). As shown in Fig. 2D, following treatment with TV, the mean fluorescent intensity of the PLA signal representing EVI1 and TBL1 interaction was significantly reduced, as also shown in the violin plot in the panel. We next interrogated by co-immunoprecipitations (Co-IPs) whether immuno-precipitated EVI1 physically bound to TBL1. Although following the pull-down of EVI1 by an anti-EVI1 antibody was unsuccessful binding with TBL1, previously reported binding of the co-repressor CtBP to EVI1 was observed (Fig. S2C) (44). This suggests that co-localization of EVI1 with TBL1 may occur as a member of the TBL1 complex without a direct binding of EVI1 to TBL1.

Figure 2. Treatment with TV blocks TBL1 co-localization with key transcription factors β-catenin, LEF1 and EVI1. A.

AML191 cells were treated with 100 nM of TV (TV) for 16 hours. Cells were cytospun onto glass slides, fixed, permeabilized and stained with anti-LEF1 and anti-TBL1 antibodies. Nuclei were stained with DAPI. Cells were imaged by confocal immunofluorescence microscopy for LEF1 and TBL1 expression. Quantification of LEF1 staining intensity was normalized to DAPI signal. Mean fluorescent intensity of 100 cells is shown. ****= p ≤ 0.0001 compared to untreated cells determined by a two-tailed, unpaired t-test. B-C. AML191 and UCSD-AML1 cells were treated with 100 nM of TV for 16 hours. Cells were cytospun onto glass slides, fixed, permeabilized and stained with anti-EVI1 and anti-TBL1 antibodies. Nuclei were stained with DAPI. Cells were imaged by confocal immunofluorescence microscopy for EVI1 and TBL1 expression. Quantification of EVI1 staining intensity was normalized to DAPI signal. Mean fluorescent intensity of 100 cells is shown. ****= p ≤ 0.0001 compared to untreated cells determined by a two-tailed, unpaired t-test. D. AML191 cells were treated with 100 nM of TV for 16 hours. Proximity ligation assay (PLA) was performed on fixed cells for TBL1 and β-catenin or EVI1 and TBL1 using the DuoLink kit from Sigma Aldrich. PLA signal was detected by confocal microscopy. Representative images of PLA signal (green) in control and TV-treated cells are shown. Mean fluorescence intensity of PLA signal for TBL1 and β-catenin (left) or EVI1 and TBL1 (right) of 25 cells per condition normalized to DAPI. ***= p ≤ 0.005 compared to untreated cells determined by a two-tailed, unpaired t-test.

EVI1 depletion is mechanistically linked to differentiation and loss of viability of human AML cells with inv(3q26) or t(3:3).

To determine whether reduction in EVI1 levels by TV is mechanistically linked to its lethal activity in AML cells, we determined the impact of CRISPR-mediated depletion of EVI1 in human AML cells with inv(3q26). For this, two separate EVI1-targeted gRNA pools were nucleofected into AML191 and UCSD-AML1 cells and compared to the control non-targeted gRNA (39). As shown in Fig. S3A, gRNA #1 (targeting exon 3 and 7) was more potent than gRNA #2 (targeting exon 1 and 2) in reducing the protein levels of EVI1, which was associated with reduction in mRNA levels of MECOM, but upregulation of the mRNA levels of ITGAM (encodes for CD11b) in AML191 cells (Fig. S3B) (42). This was also associated with a reduction in the previously described mRNA expressions representing the stemness score (Fig. S3C) (45). Consistent with this, gRNA #1 nucleofection reduced cell proliferation and induced morphologic features of differentiation as well as increased apoptosis in AML191 cells (Fig. S3D to S3F) (42). In UCSD-AML1 cells, gRNA #2 was more effective in reducing EVI1 protein levels (Fig. S4A). It also yielded similar effects on MECOM and ITGAM mRNA levels, as well as reduced the 6-gene stemness score (Fig. S4B and S4C) (45). Additionally, in UCSD-AML1 cells, gRNA #2 also reduced cell proliferation and induced differentiation (Fig. S4D and S4E). We next determined whether gRNA #1-mediated reduction in EVI1 levels would inhibit or further sensitize AML191 cells to treatment with TV or BET inhibitor. As shown in Fig. S3G, gRNA #1-mediated depletion in EVI1 significantly increased apoptosis induced by TV and OTX015. Similar partial sensitization to lethal effect of TV by nucleofection of gRNA #2 was also seen in UCSD-AML1 cells (Fig. S4F). These findings confirm that depletion of EVI1 is mechanistically linked to induction of differentiation and loss of viability of AML191 cells and sensitized them to lethality induced by TV and BETi.

Treatment with TV attenuates the activity of enhancers and reduces transcript levels of MYB, MYC and MECOM.

We next determined the effects of TV on active super enhancers (SE)/enhancers (Es) in AML cells with inv(3q26) or t(3:3). ChIP-Seq analysis to determine H3K27Ac and H3K4Me1 occupancy in AML191 cells treated for 16 hours demonstrated that TV treatment reduced the overall H3K27 peak density at SEs/Es and promoters and reduced the number of active enhancers marked with overlapping H3K27Ac and H3K4Me1 peaks (Fig. 3A, S5A and S5B). Ranked ordering of super enhancers (ROSE) plot determined by H3K27Ac ChIP-Seq analysis demonstrated loss of rankings of the SEs of HOXA9, MYB, SPI1 and GATA2 and reduced ranking of the SEs of CD34 and RARA (Fig. 3B). In contrast, the ranking of the SE of CEBPA improved, representing increased activity (Fig. 3B). Given the role of c-Myb in regulating EVI1 expression through GATA2 enhancer (46), notably, ChIP-Seq determined signal-density IGV plots of H3K27Ac at the MYB and MECOM locus also showed decline in H3K27Ac occupancy (Fig. 3C and S5C) (42). Among the EVI1 target genes, the Es and promoters of ERG, STAT1/6, SMYD3, TCF7L2 and GATA2 showed a log2 fold-decline while those of KLF6, BCL6 and LMO2 showed an increase in H3K27Ac occupancy (Fig. 3D and S5D). As expected from its inhibitory effects on nuclear β-catenin and TCF7L2 levels, ATAC-Seq analysis showed that TV treatment reduced chromatin accessibility in AML-191 cells at the WNT target genes, including LGR4/5, ASH2L, RSPO3, FZD1/8, DKK1, WNT3A, LEF1, TCF7L1/2 and XIAP (Fig. 3E). Utilizing Hi-ChIP analysis with anti-H3K27Ac antibody, we also assessed the effects of TV treatment on the looping interactions among the enhancers within the MYC SE and with the promoter. As shown in Fig. 3F, exposure of AML191 cells to 100 nM of TV significantly attenuated these interactions within the MYC SE and with MYC promoter (39). Additionally, TV also reduced the loop interactions within the BCL2 SE (Fig. S5E) (47). Utilizing the H3K27Ac ChIP-Seq findings, we also determined that treatment with TV perturbed the core transcriptional regulatory circuit (CRC) in AML-191 cells, with a marked reduction in the CRC score and depletion of MYB contribution to the CRC (Fig. S5F) (48). To determine the effects of TV on the enhancers and CRC on the transcriptome, we conducted RNA-Seq analyses following treatment of AML191 and UCSD-AML1 cells with 100 nM of TV for 16 hours. Fig. S6A and S6B show the heatmap and Venn diagram of the mRNA expression perturbations and overlapping mRNA upregulations (2262 mRNAs) or down regulations (2167 mRNAs) in tegavivint treated AML191 and UCSD-AML1 cells. TV treatment caused negative enrichment of mRNA gene-sets of MYC targets and cell cycle G2M checkpoint, while the mRNA gene-sets of TNFα signaling via NFkB, inflammatory response, interferon α/γ response, TP53 pathway and TGFβ signaling pathway were positively enriched (Fig. 4A). In AML191 and UCSD-AML1 cells, significant log2 fold-reductions in the MYC targets and cell-cycle regulatory genes but increase in apoptosis regulatory and NFkB and Inflammation pathway gene expressions were also observed in the two cell types (Fig. 4B, 4C and S6C to S6E). Notably, TV treatment also caused log2 fold-perturbations in EVI1 targets, including upregulation of GADD45B, BCL6, WNT9A, STAT1/6, CBL and SOCS1, but depletion in mRNAs of MYC, LMO2, USP2, PTGDR2 and SSN3 and reduced stemness scores (Fig. 4D, S6F and S6G). The volcano plot of mRNA expression perturbations following treatment with TV, depicted in Fig. 4E, shows concordant and significant increases in mRNA expression of SMAD7, CD86, TNFα, TRAF1, TNFSF9/15 and XIAP, but repression of mRNA of MYC, MECOM, CD34, KIT and MCM2/4 in AML191 and UCSD-AML1 cells. Taken together these findings are consistent with the observed disruptive effects of TV on β-catenin/TBL1/LEF1/EVI1 interactions that may contribute to the mRNA perturbations and the lethal activity of TV in AML cells harboring inv(3)/t(3;3) lesions and EVI1 overexpression.

Figure 3. Treatment with TV reduces H3K27Ac occupancy on super enhancer driven genes, the MECOM locus and affects chromatin looping of the MYC super enhancer with its promoter in 3q26.2-rearranged AML cells.

A. AML191 cells were treated with 100 nM of tegavivint for 16 hours. Cells were cross-linked with paraformaldehyde and H3K27Ac and H3K4Me1 ChIP-Seq analysis was performed. Peak density of H3K27Ac and H3K4Me1 occupancy in AML191 cells following treatment with 100 nM of TV (blue line) for 16 hours is shown compared to control cells (green line). B. Ranked ordering of super enhancers (ROSE) determined by H3K27Ac ChIP-Seq in AML191 cells treated with TV for 16 hours or control cells. C. IGV signal density plot of H3K27Ac ChIP-Seq peaks on the MECOM locus in AML191 following 16 hours of TV treatment compared to control cells. D. Log2 fold-change in H3K27Ac occupancy on significantly altered (p < 0.05) EVI1 target gene loci in AML191 cells treated with 100 nM of TV for 16 hours compared to control cells. E. AML191 cells were treated with 100 nM of TV for 16 hours. Cells were harvested and utilized for ATAC-Seq analysis. Graph shows log2 fold-change in chromatin accessibility at the TSS +/− 5 kb or gene body +/− 100 kb (enhancers) of significantly altered (p < 0.05) WNT target genes. F. AML191 cells were treated with 100 nM of TV for 16 hours. Cells were cross-linked with paraformaldehyde and H3K27Ac Hi-ChIP analysis was performed. Loop calling of H3K27Ac Hi-ChIP identified loops between the MYC super enhancer, outlined in orange, and the MYC promoter, highlighted in yellow that were reduced by treatment with TV. The Arima Hi-ChIP kit was used to perform this experiment and the WashU Epigenome browser was used for visualization of H3K27Ac signal density and the arc plot (chromatin loops).

Figure 4. Treatment with TV induced inflammatory response and apoptosis gene sets with concomitant depletion of MYC target gene sets and EVI1 target genes in 3q26.2-rearranged AML cells.

A. UCSD-AML1 and AML191 cells (biologic replicates) were treated with 100–200 nM of TV for 16 hours. Total RNA was isolated and utilized for RNA-Seq analysis. Graph shows selected normalized enrichment scores (NES) from gene set enrichment analysis (GSEA) of significant (p < 0.05) differentially expressed genes in UCSD-AML1 and AML191 cells compared to HALLMARK pathways. All q-values are less than 0.1. B-D. Log2 fold-changes of individual significant (p < 0.05) differentially expressed genes from the HALLMARK_MYC_TARGETS, EVI1 target genes and HALLMARK_APOPTOSIS pathways in AML191 and UCSD-AML1. E. Volcano plot of RNA-Seq determined significant, differentially expressed (1.25-fold change up or down and p < 0.05) in TV-treated over control (TV/Control) AML191 (left) or UCSD-AML1 (right) cells.

TV treatment reduces EVI1 and depletes phenotypically characterized LSCs harboring inv(3)/t(3;3).

Utilizing RPPA and CyTOF analyses, we next determined the effects of TV treatment on the protein levels in AML cells harboring inv(3)/t(3;3). Heatmap of TV-mediated, significant, log2 fold-protein level alterations are shown in Fig. 5A. Treatment with TV caused significant (p < 0.05), log2 fold-decline in the levels of p-Src, β-catenin, EVI1 and c-KIT, while increasing levels of heat shock proteins and molecular chaperones, e.g., HSP60, TRAP1, GRP78, HSP27, and p21 (Fig. 5B and 5C). CyTOF analyses conducted on AML 191 cells that were phenotypically characterized as AML LSCs, due to expression of CD34 and high expression of CD33, CD123, CLEC12A and CD99, demonstrated depletion of β-catenin TBL1/R, c-Myc, EVI1, Survivin, NOTCH1, RUNX1, as well as BCL2, MCL1 and Ki67 (Fig. 5D). However, the levels of CD11b and cleaved PARP were increased (Fig. 5D). These changes in protein expressions were associated with reduction in the % of AML191 and AML194 cells co-expressing CD34 and CD117 (Fig. 5E and 5F). CyTOF analyses of cells from three additional AML samples demonstrated that TV treatment reduced expression of CD34, CD99, CD123, CD244 and CLEC12A, but increased the expression of CD11b and cleaved PARP in each of the cell-types with inv(3)/t(3;3) (Fig. 5G). Collectively, these findings confirm that treatment with TV reduces the levels of EVI1, c-Myc and the key anti-apoptotic proteins, while inducing the expression of differentiation and apoptosis marker in phenotypically defined LSCs.

Figure 5. TV treatment depletes EVI1 and MYC levels in phenotypically defined leukemia stem cells.

A. AML191 cells were treated with 50 nM of TV for 24 hours. Cells were harvested and utilized for Reverse Phase Protein Array (RPPA) analysis. The heat map shows the differentially expressed proteins due to TV treatment compared to untreated control B-C. Selected significant (p < 0.05) differentially expressed proteins (decreased [B] or increased [C]) by TV treatment compared control. D. PD AML191 cells were treated with 100 nM of TV for 16 hours. Cells were fixed with 4% paraformaldehyde and utilized for Cytometry by Time of Flight (CyTOF) analysis. Analysis was performed utilizing Astrolabe software. Heat map shows the absolute fold-change in gene expression following TV treatment compared to DMSO control in phenotypically defined AML stem cells. E-F. AML191 and AML194 cells were treated with the indicated concentrations of TV for 16 hours. Live cells were stained with anti-CD34 and anti-CD117 antibodies and analyzed by flow cytometry. The panels show the % CD34 and CD117 double-positive AML191 (E) or AML194 (F) cells following treatment with TV compared to the control cells. Columns are the mean of three independent experiments run in duplicate; Bars, represent the S.E.M. * =p ≤ 0.01, **=p ≤ 0.001, ***=p ≤ 0.0001. G. Fold change in gene expression of stem/progenitor cell markers and selected proteins determined by CyTOF in three patient-derived 3q26-rearranged AML samples following treatment with 100 nM of TV for 16 hours. Analysis was performed utilizing Astrolabe software.

CRISPR screen determined dependency on BRD4 and synergistic lethality of co-treatment with TV and BETi AML in cell lines and PD AML cells harboring 3q26.2 lesions.

We next conducted, in UCSD-AML1 transduced with Cas9, a CRISPR screen with a previously reported, targeted, domain-specific gRNAs library against epigenetic regulators (42). The screen revealed, among other ‘druggable’ targets, BRD4 as a dependency (Fig. S7A and S7B). Guided by this and based on the observation that TV treatment reduced SE-driven c-Myc and BCL2 levels in LSCs (Fig. 5D), we next determined whether co-treatment with TV and BETi would exert superior lethal activity in AML cells with inv(3)/t(3;3). Fig. 6A and 6B show that co-treatment with TV and OTX015 synergistically induced apoptosis in AML191 cells, with delta synergy scores of >1 by the ZIP method. This was associated with greater reduction in protein levels (by immunoblot analysis), in cell lysates, of EVI1, c-Myc, c-Myb, RUNX1, PU.1, CDK6 and c-KIT, as well as in anti-apoptotic proteins MCL-1, BCL2, and Bcl-xL, with upregulation of BIM, p21 and p27, in AML191 cells (Fig. 6C). Combined treatment with relatively lower concentrations of TV (20 to 50 nM) and OTX015 (250 to 500 nM) also yielded delta synergy scores of > 1 in AML194, UCSD-AML1, OCI-AML20, MUTZ3 and HNT34 cells (Fig. 6D). Notably, in 6 samples of PD AML cells with inv(3)/t(3;3), co-treatment with 30 nM or less of TV and OTX015 (100 to 500 nM) also synergistically induced apoptosis (Fig. 6E). In contrast, treatment of three normal CD34+ hematopoietic progenitor cells (HPCs) with TV or OTX015 alone induced less than 15% loss of viability, and their co-treatment exerted sub-additive lethal effects (Fig. S7C). We next determined the effect of TV and OTX015 on AML cells harboring atypical MECOM rearrangements, e.g., an AML sample with t(3;8) (q26;q24), involving the MYC SE, or two AML samples with t(3;21) (q26;q22), involving the RUNX1 locus, as well as other co-mutations (Fig. S7D) (17). These translocations involving chromosome 3q26.2 have been previously reported to exhibit EVI1 overexpression (17). Treatment of all three AML cell-types with OTX015 (100 to 500 nM) or TV (3 to 30 nM) showed modest loss of viability (Figs. S7E and S7F). Additionally, co-treatment with TV and OTX015 synergistically induced apoptosis in AML cells with t(3,8) (q26;q24) (Fig. S7G). We next determined the single agent lethal activity of venetoclax against AML cells with inv(3)/t(3;3). As shown, monotherapy with venetoclax dose-dependently induced apoptosis in most AML cell lines and PD AML cells with inv(3)/t(3;3) (Fig. S7H and S7I). Treatment with venetoclax also dose-dependently induced loss of viability in PD AML cells lacking genetic alterations involving chromosome 3q26.2 (Fig. S7I). Co-treatment with TV and venetoclax also synergistically induced apoptosis, with delta synergy scores of >1 by the ZIP method, in the AML cell lines, as well as in 3 samples of PD AML cells with inv(3)/t(3;3) (Fig. 6F and 6G). In contrast, treatment of three normal CD34+ hematopoietic progenitor cells (HPCs) with TV or venetoclax alone induced less than 15% loss of viability, and their co-treatment exerted sub-additive lethal effects (Fig. S7J). We also determined whether co-treatment with TV and the histone acetyltransferase (HAT) CBP/p300 inhibitor GNE-049 would also exert synergistic in vitro lethality in AML cells with inv(3)/t(3;3)(49). As shown in Fig. S7K and S7L, combined therapy with the two agents induced synergistic lethality. We next determined whether co-treatment with BETi and venetoclax would exert synergistic lethality against AML cell lines or PD AML cells with inv(3)/t(3;3). Fig. S7M to S7P demonstrate that combined therapy with OTX015 and venetoclax synergistically induced apoptosis in AML191, UCSD-AML1 cells and AML194, as well as in 5 samples of PD AML cells with inv(3)/t(3;3). The lethal activity of OTX015 and venetoclax was further enhanced by combination with tegavivint in UCSD-AML1 and AML191 cells (Fig. S7Q and S7R).

Figure 6. Co-treatment with TV and OTX015 or venetoclax exerts synergistic lethal activity against 3q26-rearranged AML cells.

A. AML191 cells were treated with the indicated concentrations of TV and/or OTX015 for 48 hours. The % of annexin V-positive, apoptotic cells were determined by flow cytometry. Values are the mean of three or more independent experiments. B. Delta synergy scores for combination of TV and OTX015 in AML191 cells. Synergy was determined utilizing the SynergyFinder V3 web application. Delta synergy scores (ZIP method) greater than 1.0 indicate a synergistic interaction of the two agents. C. AML191 cells were treated with the indicated concentrations of TV and/or OTX015 for 24 hours. At the end of treatment, total cell lysates were prepared and immunoblot analyses were conducted. The expression of β-Actin in the lysates served as the loading control. D. Table of delta synergy scores (by ZIP method) based on average percent apoptosis values of TV and OTX015 combinations in 3q26-rearranged AML cell lines AML191, AML194, UCSD-AML1, OCI-AML20, MUTZ3 and HNT34. E. PD 3q26-rearranged AML samples (n=6) were treated with the indicated concentrations of TV and/or OTX015 for 48 hours. The % of TO-PRO-3 iodide-positive, non-viable cells were determined by flow cytometry. Delta synergy scores were determined utilizing the SynergyFinder V3 web application. Delta synergy scores (ZIP method) greater than 1.0 indicate a synergistic interaction of the two agents. F. Table of delta synergy scores (by ZIP method) based on average percent apoptosis values of TV and venetoclax combinations in 3q26 rearranged AML cell lines AML191, AML194, UCSD-AML1, OCI-AML20, MUTZ3 and HNT34. G. PD 3q26-rearranged AML samples (n=3) were treated with the indicated concentrations of TV and/or venetoclax for 48 hours. The % TO-PRO-3 iodide-positive, non-viable cells were determined by flow cytometry. Delta synergy scores were determined utilizing the SynergyFinder V3 web application. Delta synergy scores (ZIP method) greater than 1.0 indicate a synergistic interaction of the two agents.

Superior in vivo efficacy of combined therapy with tegavivint and BET inhibitor or venetoclax against AML with 3q26 lesions.

We next determined the in vivo anti-leukemia efficacy of TV and/or OTX015 in aggressive, flank or tail-vein infused, xenograft models of AML cells with inv(3)/t(3;3) in immune-depleted NSG mice. These models were transduced with Luciferase/GFP for bioluminescence imaging. First, cohorts of mice were injected in the flank with AML 191 cells (in a 1:1 suspension of Matrigel) and tumors were measured once a week, beginning 3 weeks after injection. Mice were treated for 4 weeks with vehicle control, or with TV or OTX015 alone, or co-treated with TV and OTX015. The dose of each drug employed here was previously determined to be safe (18,23,47). Whereas monotherapy with TV delayed tumor growth, co-treatment with TV and OTX015 was significantly superior in delaying and suppressing AML growth compared to vehicle or each drug alone (Fig. 7A and S8). Consistent with this, co-treatment with TV and OTX015 was significantly superior in improving the survival of the mice up to > 1.5 cm of tumor growth, when mice had to be euthanized (Fig. 7B). Indeed, co-treatment with TV and OTX015 yielded a plateau in the survival curve for up to 70 days with 100% of mice in the cohort surviving, compared to none in the other cohorts (p < 0.01) (Fig. 7B). Although the tail-vein infused model of AML191 cells was rapidly lethal, the effects on AML burden could still be assessed and showed significantly greater reduction in AML burden, following co-treatment with TV and OTX015 compared to each drug alone (Fig. 7C). We next determined the anti-AML efficacy of TV and/or OTX015 against a separate, tail vein infused and engrafted, aggressive xenograft model of PD AML242 cells with inv(3q26) (Table S1). Since vehicle treated control mice engrafted with AML242 had to be euthanized within 3 weeks of starting drug treatments, effects on AML burden were imaged and evaluated after only 2-weeks of the drug treatments. As shown in Fig. 7D, treatment with TV or OTX015 significantly reduced AML burden, and co-treatment with TV and OTX015 was even more effective. Compared to treatment with vehicle, TV or OTX015 alone, combined therapy with the two agents significantly improved survival of the NSG mice engrafted with AML242 cells (Fig. 7E). We next determined the anti-AML efficacy of TV and/or venetoclax against the flank engrafted xenograft model of AML191 cells. Fig. 7F demonstrates that 4-week treatment with either agent alone or in combination delayed AML growth. In contrast, compared to monotherapy with each drug alone or vehicle control, co-treatment with the two drugs modestly but significantly improved survival of the NSG mice (Fig. 7G). Additionally, combined therapy with TV and venetoclax also significantly reduced AML burden in the tail-vein infused xenograft model of AML191 cells (Fig. 7H). These findings indicate that co-treatment with TV and OTX015 or venetoclax are effective combination therapies worthy of further in vivo testing and development in AML with inv(3)/t(3;3).

Figure 7. Compared to single agent treatment, co-treatment with TV and OTX015 or venetoclax exerts superior anti-leukemia activity and improves survival of mice bearing 3q26.2-rearranged AML xenografts.

A. One million luciferase-expressing AML191 cells suspended in Matrigel were injected into the flanks of five 4–6 week old male NSG (NOD, scid, gamma) mice per condition. Tumor measurements and treatment started 21 days post cell injection. Mice were randomized to equivalent mean tumor volume prior to starting treatment. Mice were treated with vehicle, TV (50 mg/kg twice weekly, by intraperitoneal injection), and/or OTX015 (30 mg/kg, daily 5 days a week by oral gavage) for 4 weeks. Tumors were measured once per week until tumor volumes reached 1.5 cm³. Treated mice were followed until ~70 days post injection. B. Percent survival of AML tumor bearing mice from A. Survival was determined by tumor growth less than 1.5 cm³. C. Following a pre-conditioning dose of irradiation, 4–6 week old female NSG mice were injected in the lateral tail vein with four million luciferase-expressing AML191 cells and monitored for seven days. Mice were imaged by Xenogen camera and randomized to equivalent bioluminescence. Mice were treated with vehicle, TV (50 mg/kg twice weekly, by intraperitoneal injection), and/or OTX015 (30 mg/kg, daily 5 days a week by oral gavage) for 5 weeks. Mice were imaged by Xenogen camera to determine treatment efficacy. Total bioluminescent flux for each condition is reported in photons/second [p/s]. * p ≤ 0.01, ** p ≤ 0.001, *** p ≤ 0.0001. D. One and a half million luciferase-expressing PDX AML242 cells were injected into the tail vein of non-lethally irradiated female NSG mice (4–6 weeks old). Seven days post infusion, mice were imaged by Xenogen camera and randomized to equivalent bioluminescence. Mice were treated with vehicle, TV (50 mg/kg twice weekly, by intraperitoneal injection), and/or OTX015 (30 mg/kg, daily 5 days a week by oral gavage) for 2 weeks. Mice were imaged by Xenogen camera to determine treatment efficacy. Total bioluminescent flux for each condition is reported in photons/second [p/s]. * p ≤ 0.01, ** p ≤ 0.001, *** p ≤ 0.0001. E. Kaplan-Meier survival curves for NSG mice bearing luciferase-expressing PDX AML242 xenografts treated with TV and/or OTX015 for 3 weeks. Significance was calculated with a Mantel-Cox log rank test and p < 0.05 are considered significant. * p ≤ 0.01, ** p ≤ 0.001, *** p ≤ 0.0001. F. Male NSG mice were injected with luciferase-expressing AML191 cells as in A. When tumors reached ~100 mm³, mice were randomized and treated with vehicle, TV (50 mg/kg twice weekly, by intraperitoneal injection), and/or venetoclax (30 mg/kg, daily 5 days a week by oral gavage) for 4 weeks. Tumors were measured once per week until tumor volumes reached 1.5 cm³. Treated mice were followed until ~70 days post-injection. G. Percent survival of AML tumor-bearing mice treated with TV and/or venetoclax as indicated in F. Survival was determined by tumor growth less than 1.5 cm³. H. Following a pre-conditioning dose of irradiation, 4–6 week old female NSG mice were injected in the lateral tail vein with four million luciferase-expressing AML191 cells and monitored for seven days. Mice were imaged by Xenogen camera and randomized to equivalent bioluminescence. Mice were treated with vehicle, TV (50 mg/kg twice weekly, by intraperitoneal injection), and/or venetoclax (30 mg/kg, daily 5 days a week by oral gavage) for 5 weeks. Mice were imaged by Xenogen camera to determine treatment efficacy. Total bioluminescent flux for each condition is reported in photons/second [p/s]. * p ≤ 0.01, ** p ≤ 0.001, *** p ≤ 0.0001.

Discussion:

Findings presented here demonstrate that treatment with a pan-BETi or TV induces loss of viability in cultured cell lines and PD AML cells harboring chromosomal alterations inv(3)/t(3;3). BETi treatment repressed enhancer driven and AML relevant oncoproteins, including EVI1, c-Myc, c-Myb and MCL1, while inducing HEXIM1 levels. For the first time, our findings also show that TV-induced lethality is associated with reduced interaction of TBL1 with LEF1 and EVI1 (although the interaction of EVI1 with TBL1 may occur through a member of the TBL1 complex without the direct binding of EVI1 to TBL1), in addition to depletion of nuclear β-catenin, resulting in repression of EVI1, c-Myc, c-Myb, BCL2/Bcl-xL/MCL1. Exposure to TV also repressed ERG1, which was recently shown to be a direct target of EVI1 and contributes to enhanced cell survival and differentiation arrest of AML cells with inv(3)/t(3;3) and EVI1 overexpression (27). Consistent with these observations, our findings also demonstrate that co-treatment with a BETi and TV induced synergistic in vitro lethality, as well as significantly reduced AML burden and improved survival of NSG mice engrafted with AML cells with inv(3)/t(3;3). Also shown here is that TV mediated reduction in EVI1 levels is mechanistically responsible for the synergistic lethal activity with BETi, since partial depletion of EVI1 via CRISPR-gRNA induced differentiation and loss of viability and promoted increased lethal activity of TV in AML cells harboring inv(3)/t(3;3).

Consistent with TV-mediated repression of multiple transcriptional regulators in AML cells, including EVI1, TV treatment inhibits enhancer activities at several loci including MECOM, GATA2, MYC, MYB, EGR and BCL2 but upregulated BCL6 enhancer activity (47). Notably, Hi-ChIP with anti-H3K27Ac also revealed inhibition of enhancer interactions at MYC and BCL2 loci. TV-induced marked reduction in the core transcriptional regulatory circuit also explained its wide-ranging effects on the transcriptional gene-sets determined by RNA-Seq analysis. These included negative enrichment of gene-sets involving MYC, E2F and cell-cycle regulation and positive enrichment of gene-sets of inflammatory response, interferon α/γ response, TNFα signaling via NFkB, TGFβ and STAT3/5 signaling, as well as TP53 and apoptosis pathways. Among those downregulated were mRNA expressions of MECOM, MYC, MYB and MCM2/4 and upregulated expressions of TNFα, TRAF1, SMAD7 and XIAP in AML cells harboring inv(3). Because MYB binds to a specific site at the relocated GATA2 enhancer in AML cells harboring inv(3)/t(3;3) to upregulate EVI1 (46), repression of MYB by TV also likely contributes to TV-induced repression of EVI1. Together with repression of LEF1, ERG, TV-mediated repression of MYC and MYB, collectively, these transcriptional alterations also explain TV-induced reduction in the AML stemness score (45). They also provide an explanation for the reduction in protein expressions of EVI1 and c-Myc and in those that phenotypically characterize AML initiating LSCs via CyTOF analyses, resulting in reduction of the % of AML LSCs that express CD34 and CD117.

It is noteworthy that the gRNAs for BRD4 and EP300 were enriched in a CRISPR screen in AML cells with inv(3) involving 3q26, while partial depletion of EVI1 via CRISPR knockout sensitized AML cells to BETi. These observations, coupled with the findings that in AML cells harboring inv(3) and EVI1 overexpression treatment with BETi, as well as with TV, reduced EVI1, c-Myb, c-Myc, CDK4/6 and MCL1, support and explain why co-treatment with BETi and TV exerts synergistic in vitro lethal activity and in vivo efficacy in models of AML cells harboring inv(3)/t(3;3). This combination also induced synergistic in vitro lethality in AML cells with atypical MECOM locus translocations, including t(3;8) and t(3;21), which are also associated with EVI1 and c-Myc expression (17). Moreover, since the HAT CBP/p300 is also involved in MYB regulated EVI1 expression (46), the CRISPR screen findings also explain the synergistic lethality we observed here in AML cells with inv(3)/t(3;3) due to co-treatment with TV and the CBP/p300 inhibitor GNE-049. Resistance to the BCL2 inhibitor venetoclax in AML setting has been attributed to upregulation of MCL1 and Bcl-xL following treatment with venetoclax or venetoclax-based anti-AML therapies (50). Treatment with TV and a pan-BET inhibitor not only repressed EVI1 and c-Myc levels but also reduced protein levels of MCL1 and Bcl-xL contributing to the observed in vitro synergistic lethality and superior in vivo efficacy of co-treatment with TV and venetoclax in AML cells harboring inv(3)/t(3;3). This is schematically represented in Fig. S9. The marked reduction in EVI1, c-Myc, c-Myb and MCL-1 due to BETi treatment also explains the synergistic lethal activity of co-treatment with BETi and venetoclax presented here in AML cells with inv(3)/t(3;3) (Fig. S9). This raises the possibility that a triple drug combination involving BETi, TV and venetoclax may exert, if safe, even more potent activity against this high-risk AML subtype that exhibits poor clinical outcome following standard AML therapies, including other venetoclax-based drug regimens (26).

If past clinical experience with targeted therapies that have yielded incremental improvement in clinical outcome in AML is to be the guide, then findings presented here underscore that, at safe doses further in vivo interrogation of the combination of TV and BETi, with or without venetoclax, in AML with inv(3)/t(3;3) is warranted (50). AML-initiating, stem/progenitor cells have been reported to be enriched in the minimal and measurable residual disease (MRD) state after achieving a complete remission of AML (51). Therefore, findings presented here are promising that, in AML harboring inv(3)/t(3;3) with EVI1 overexpression, TV treatment attenuates the core transcriptional regulatory circuitry, reduces AML stemness score and depletes phenotypically characterized AML stem/progenitor cells. These findings also strongly support future development and in vivo testing of TV-based combinations highlighted here in the clinical setting of AML with inv(3)/t(3;3) in MRD positive clinical remission.