Functional diversity of inhibitors tackling the differentiation blockage of MLL-rearranged leukemia

Abstract

Introduction: The chromosomal rearrangements of the mixed-lineage leukemia gene MLL (KMT2A) have been extensively characterized as a potent oncogenic driver in leukemia. For its oncogenic function, most MLL-fusion proteins exploit the multienzyme super elongation complex leading to elevated expression of MLL target genes. High expression of MLL target genes overwrites the normal hematopoietic differentiation program, resulting in undifferentiated blasts characterized by the capacity to self-renew. Although extensive resources devoted to increased understanding of therapeutic targets to overcome de-differentiation in ALL/AML, the inter-dependencies of targets are still not well described. The majority of inhibitors potentially interfering with MLL-fusion protein driven transformation have been characterized in individual studies, which so far hindered their direct cross-comparison.

Methods: In our study, we characterized head-to-head clinical stage inhibitors for BET, DHODH, DOT1L as well as two novel inhibitors for CDK9 and the Menin-MLL interaction with a focus on differentiation induction. We profiled those inhibitors for global gene expression effects in a large cell line panel and examined cellular responses such as inhibition of proliferation, apoptosis induction, cell cycle arrest, surface marker expression, morphological phenotype changes, and phagocytosis as functional differentiation readout. We also verified the combination potential of those inhibitors on proliferation and differentiation level.

Results: Our analysis revealed significant differences in differentiation induction and in modulating MLL-fusion target gene expression. We observed Menin-MLL and DOT1L inhibitors act very specifically on MLL-fused leukemia cell lines, whereas inhibitors of BET, DHODH and P-TEFb have strong effects beyond MLL-fusions. Significant differentiation effects were detected for Menin-MLL, DOT1L, and DHODH inhibitors, whereas BET and CDK9 inhibitors primarily induced apoptosis in AML/ALL cancer models. For the first time, we explored combination potential of the abovementioned inhibitors with regards to overcoming the differentiation blockage.

Conclusion: Our findings show substantial diversity in the molecular activities of those inhibitors and provide valuable insights into the further developmental potential as single agents or in combinations in MLL-fused leukemia.

Keywords: Acute lymphoblastic leukemia (ALL); Acute myeloid leukemia (AML); BET; DHODH; DOT1L; MLL-fusion; Menin-MLL; P-TEFb; Small molecule inhibitors.

Fig. 1

Comparing the inhibitors tackling the differentiation blockage in AML. a Proliferation inhibition effect of BAY-155, OTX015, EPZ-5676, BAY 1251152, and Brequinar in MV4-11, MOLM-13 and HL-60 cells after 4 days of treatment. Vehicle control (DMSO) of each cell line is set to 100%. Representative experiment of at least 3 biological replicates is shown. b Quantification of CD11b expression after BAY-155, OTX015, EPZ-5676, BAY 1251152, and Brequinar in MV4-11, MOLM-13, and HL-60 cell lines after 4 and 7 days of treatment detected with flow cytometry. Data represents median values from 10,000 living cells normalized to vehicle control (DMSO). c Wright-Giemsa-stained cytospins of MV4-11, MOLM-13, and HL-60 cells after 7 days of treatment with BAY-155 (0.05 μM, 0.5 μM, 2.5 μM, respectively), OTX015 (0.1 μM, 0.1 μM, 0.25 μM, respectively), EPZ-5676 (0.4 μM, 1 μM, 10 μM, respectively), BAY 1251152 (0.05 μM, 0.05 μM, 0.05 μM, respectively), and Brequinar (0.1 μM, 0.3 μM, 1 μM, respectively). In the top right corner of each image magnification of representative cells are shown. Black scale bar indicates 10 μm

Fig. 2

Inhibitor-induced differential gene expression and anti-proliferative effects in AML/ALL cells. a IC₅₀ (μM) values of BAY-155, OTX015, EPZ-5676, BAY 1251152, and Brequinar after 1, 4, and 7 days of treatment. In the case where IC₅₀ determination was beyond tested maximal concentration (10 μM), > 10 μM is used. ALL cell lines are indicated by gray name box shading. b Number of genes upregulated (red bars) and downregulated (blue bars) upon treatment by the indicated inhibitors in non-fused MLL-WT (left from the dashed line) and MLL-fused (right from the dashed line) models (log2FC > 1, FDR < 0.1). pIC₅₀ (− log10 (IC₅₀) in M) values (diamond shape) for the indicated inhibitors and cell models after 4 days of treatment. In the situation where pIC₅₀ determination was beyond maximal concentration, no data point is shown. c Correlation of differential gene expression effects between inhibitors and cell models. Heatmap represents correlation of log2FC of gene expression grouped based on the inhibitor in all used cell lines ordered identically to b. Black boxes indicate cell line comparison for a single inhibitor

Fig. 3

Gene set enrichment and principal component analyses. a Heatmap representing enrichment scores (NES) of different gene sets in GSEA. Yellow star indicates significant enrichment. b Analysis of logFC gene expression of indicated MLL target genes. Data is normalized to corresponding DMSO control and cell models are grouped based on their respective IC₅₀ values. c Scores plot of the PCA based on the top thousand most variable genes in AML cell lines. Data adjusted to gene expression in vehicle (DMSO). d Loadings plot corresponding to the plot in c. Genes associated with AML differentiation are highlighted

Fig. 4

Phagocytosis of E.coli particles after inhibitor-induced differentiation. a qRT-PCR analysis of representative surface markers and genes associated with hematopoietic differentiation in MOLM-13 cells after 7 days of treatment with BAY-155 (3 μM), OTX015 (0.2 μM), EPZ-5676 (6 μM), BAY 1251152 (0.05 μM), and Brequinar (3 μM). Data presented is an average of three biological replicates normalized to vehicle control (DMSO). n.e. not expressed. b Flow cytometry scatter plots (1st and 3rd left column) showing the distribution of CD11b (APC) and E.coli (FITC labeled) staining of viable MOLM-13 cells after 7 days of treatment with BAY-155, (3 μM) OTX015 (0.2 μM), EPZ-5676 (6 μM), BAY 1251152 (0.05 μM), and Brequinar (3 μM). Representative experiment of three biological replicates is shown. Immunofluorescence staining (2nd and 4th left column) of MOLM-13 cells after 7 days of treatment with BAY-155, (3 μM) OTX015 (0.2 μM), EPZ-5676 (6 μM), BAY 1251152 (0.05 μM), and Brequinar (3 μM). Representative merge image of CD11b (red), E.coli particles (yellow), and nucleus (blue) is shown

Fig. 5

Analysis of combinations effects on proliferation and CD11b expression levels. a IC₅₀-based isobologram analysis of inhibitor combinations in MOLM-13 cells after 4 days of treatment. The diagonal lines indicate additivity. Experimental data points, represented by dots (square, triangle, and circle) indicate biological replicates. b CD11b expression level after BAY-155 (0.15 μM)–Brequinar (0.64 μM), BAY-155 (0.15 μM)–EPZ-5676 (0.64 μM), Brequinar (0.15 μM)–EPZ-5676 (0.64 μM), and Brequinar (0.15 μM)–OTX015 (0.15 μM) combinations in MOLM-13 cells after 4 days of treatment. Data shows a representative concentration. Bar charts show an average of three biological replicates, the orange line indicates predicted additive effect, error shows SD, statistics ***P < 0.001, **P < 0.01, *P < 0.05, n.s. P < 0.05, two-sided t test. c Graphical summary