The T-ALL cell lines (27–29) MOLT-4 (ATCC no. CRL-1582), CCRF-CEM (ATCC no. CCL-119), P12-ICHIKAWA (DSMZ no. ACC-34), and RPMI-8402 (DSMZ no. ACC-290) were obtained from ATCC (Manassas, VA, USA) and DSMZ (Braunschweig, Germany). The cell lines have MYB-activation by gene duplication (7, 10). Cell identity was confirmed by STR-analysis and all cell lines were shown to be mycoplasma free prior to experiments. T-ALL cells were maintained in RPMI-1640 medium with GlutaMAX, 10% or 20% FBS, and 1% penicillin-streptomycin (Thermo Fisher Scientific, Waltham, MA, USA). Human mononuclear cells (Merck, Darmstadt, Germany) isolated from a healthy donor (HMCs) were maintained in Mononuclear cell medium (Merck) according to the instructions of the supplier. All cells were kept in a humidified incubator at 37°C and 5% CO₂.

The L1100 Inhibitor Library, bardoxolone methyl (CDDO-Me, RTA 402), omaveloxolone (RTA 408), and doxorubicin were purchased from Selleck Chemicals (Houston, TX, USA). The IKK inhibitor VII and IKK-2 inhibitor IV were obtained from Merck. The KI696 peptide (30) was a kind gift from Dr. Volkan Sayin. Drugs were kept at -80°C or -20°C in stock solutions of 10 µM in DMSO and diluted to working concentrations in growth medium prior to experiments. For drug screening, 4,000 MOLT-4 or CCRF-CEM cells were seeded in V-bottom 96-well plates (Greiner Bio-One, Kremsmünster, Austria) and after 24 h treated with 1 µM of 768 small-molecule inhibitors from the Inhibitor Library ( Supplementary Table S1 ). Cells were incubated with drugs, or DMSO as control, for 48 h whereafter they were washed with PBS and harvested using direct lysis (31) in 100 µl of 1 mg/ml BSA (Thermo Fisher Scientific). The screening was performed three times and cell plates were stored at -80°C until further analysis. For validation experiments, 2-2.5x10⁵ T-ALL cells were seeded in 24-well plates (RNA analysis) and 4x10⁵ cells were seeded in 6-well plates (protein analysis) and on the next day treated for 24 h with drugs or equal volumes of DMSO.

Total RNA was isolated using the RNeasy Micro-kit (Qiagen, Hilden, Germany) following the protocol of the manufacturer. RNA purity and concentration was measured with the NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific). Reverse transcription was performed with the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA, USA) according to instructions of the manufacturer. For drug screening, 2 µl of cell lysate were used in 10 µl cDNA reactions.

Total protein was isolated using RIPA buffer supplemented with the Halt protease and phosphatase inhibitor (Thermo Fisher Scientific). Protein concentrations were measured with the DC protein assay (Bio-Rad). Protein expression was analyzed by Western blotting using the NuPAGE system (Thermo Fisher Scientific) with monoclonal antibodies to MYB (clone 1-1; Merck Millipore) and beta-actin (ab8226, Abcam, Cambridge, UK). Bands were visualized with horseradish peroxidase-conjugated secondary antibodies and chemiluminescent detection with the Supersignal West Femto Max Sensitivity Substrate (Thermo Fisher Scientific). Blots were scanned with an Amersham ImageQuant 800 imaging system (Cytiva, Marlborough, MA, USA). MYB protein expression was quantified using ImageJ v2.0.0-rc-43 and normalized with beta-actin expression.

qPCR was done with the AB 7500 Fast Real-Time PCR system using TaqMan gene expression assays (Thermo Fisher Scientific) for MYB (Hs00920556_m1), CD34 (Hs00156373_m1), CCNB1 (Hs00259126_m1), MCM4 (Hs00381533_m1), CDK1 (Hs00364293_m1), and the reference genes UBC (Hs01871556_s1) and 18S (Hs99999901_s1). Relative gene expression was calculated by the ddCT method (32).

RNA-seq was performed at Eurofins Genomics (Luxembourg City, Luxembourg) using the INVIEW Transcriptome Discover service. Library preparation was done with poly A enrichment, random primed strand-specific cDNA synthesis, adapter ligation, and adapter-specific PCR amplification. Libraries were subjected to Illumina paired-end sequencing (2x150 bps) with at least 30 million read pairs generated. Gene counts were extracted from FastQ files on the Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX) cluster with FastQC/0.11.5, STAR v2.5.3a, Samtools v1.5, and FeatureCounts (Subread v1.5.2). Samples were analyzed with unsupervised hierarchical clustering using the hclust function in R and heatmaps were generated using R or Morpheus (Broad Institute). Differential gene expression was analyzed with DESeq2 v1.22.2. Gene ontology analysis was performed with ToppGene (33) and Gene set enrichment analysis (GSEA) (34) was done with GSEA v4.2.2 (Broad Institute).

For cell viability assays, 2,000 T-ALL cells were seeded in black 96-well plates (BD, Franklin Lakes, NJ, USA) and after 24 h treated with single or combinations of drugs. The AlamarBlue reagent (Thermo Fisher Scientific) was added to wells after 72 h and plates were analyzed with the VICTOR3 multilabel plate reader (PerkinElmer, Waltham, MA, USA). For apoptosis assays, 2,000 cells were seeded in Corning white 96-well plates (Corning, New York, USA) and treated with drugs after 24 h. The Caspase-Glo 3/7 Assay (Promega, Madison, WI, USA) was used to estimate apoptosis and luminescent signals were quantified on the VICTOR3. For cell cycle analysis, 8x10⁵ MOLT-4 cells were seeded in 6-well plates and on the next day treated for 24 h with drugs. Cells were supplemented with 10 μM BrdU during the last 4 h and then fixed and stained using the FITC BrdU Flow kit (BD). Stained cells were analyzed with an Accuri C6 flow cytometer (BD). All experiments were performed three times.

MYB siRNA-mediated knockdown was done with the Cell Line Nucleofector Kit L on the Amaxa Nucleofector II device (Cologne, Germany) using 2x10⁶ MOLT-4 cells according to the instructions provided by the manufacturer. Cells were electroporated with 1 µM Silencer Select MYB (s9109, s9110) or negative control siRNAs (Thermo Fisher Scientific). Total RNA and protein were isolated from transfected cells after 48 h. All experiments were done three times.

Unsupervised hierarchical clustering was done with R and overlap of differentially expressed genes was visualized with BioVenn (https://www.biovenn.nl) and analyzed with chi-square tests. Drug kernel density estimation plots were generated with the density function of the stats package of R (https://www.r-project.org). Correlation of treatment effects between T-ALL cell lines was estimated by Pearson correlation. Dose-response curves were generated and analyzed by nonlinear regression in Prism 9 (GraphPad Software, San Diego, CA, USA); IC₂₅ and IC₅₀ were defined as drug concentrations that reduced cell viability by 25% and 50%, respectively. For combination treatments, the definition of drug additivity and synergy was according to Bliss (35). Differences between experimentally measured effects of drug combinations and the expected effects calculated from single drug treatments were analyzed using independent samples t-test. Bar graph values are presented as mean ± SEM. Differences between groups in qPCR, cell cycle, and apoptosis analyses were evaluated with one-way ANOVA or independent samples t-test. All statistical tests were two-sided. A P-value of less than 0.05 was considered statistically significant.

We performed a large-scale in vitro screening of two T-ALL cell lines, CCRF-CEM and MOLT-4, with MYB-activation using 768 small-molecule drugs from the L1100 Inhibitor Library ( Figure 1 , Supplementary Table S1 ). Cells were treated with 1 μM of inhibitors for 48 h after which relative MYB mRNA levels were analyzed by qPCR. Density plots of MYB gene expression following drug treatments showed similar distributions in three independent experiments for each cell line ( Figure 1A ). There was a significant (P < 0.0001) correlation between drug effects in the two cell lines ( Figure 1B ). A larger number of drugs reduced MYB mRNA levels in CCRF-CEM cells compared with MOLT-4 cells ( Figure 1A ). Further analysis showed that 46 drugs caused a greater than 50% decrease in MYB gene expression in both cases ( Figure 1C ). These agents included histone-deacetylase inhibitors (HDACi), intracellular kinase (IK) inhibitors, receptor tyrosine kinase (RTK) inhibitors, and other drugs with diverse targets ( Figure 1D , Supplementary Table S2 ). Among these were drugs that were previously identified as inhibitors of MYB gene expression (AT7519, flavopiridol, givinostat, and vorinostat) (36–39) as well as an inhibitor of MYB protein activity (celastrol) (12). To validate our drug screen, we treated MOLT-4 cells with two of the identified HDACi (givinostat and panobinostat) and two of the cyclin-dependent kinase inhibitors (CDKi) (AT7519 and SNS-032) from the screen. All tested drugs significantly decreased MYB mRNA levels at nanomolar concentrations ( Supplementary Figure S1 ). We next used hierarchical clustering and heat map analysis of MYB mRNA levels to investigate the potency of the inhibitors of MYB gene expression identified in our screen ( Figure 1E ). These analyses identified bardoxolone methyl and omaveloxolone, two related oleanane triterpenoids (OTs), as top hits in the screen.

To confirm the effects of OT treatment on MYB gene expression in T-ALL cells, we treated MOLT-4, CCRF-CEM, P12-ICHIKAWA, and RPMI-8402 cells with nanomolar concentrations of bardoxolone methyl or omaveloxolone for 24 h. This treatment led to a dose-dependent decrease in MYB mRNA levels in all T-ALL cell lines tested ( Figure 2A , Supplementary Figure S2 ). A corresponding reduction in MYB protein levels after OT treatment was confirmed by immunoblotting ( Figure 2B , Supplementary Figure S3A ). To investigate the biological consequences of MYB suppression in OT treated cells, we analyzed the expression of known MYB target genes involved in cell cycle regulation and leukemic stem cell function. The expression of CCNB1, CDK1, MCM4, and CD34 were significantly downregulated in MOLT-4 cells with OT treatment ( Figure 2C ). To verify that these genes are regulated by MYB in T-ALL cells, we knocked down MYB mRNA and protein levels in MOLT-4 cells using siRNAs ( Figures 2D, E , Supplementary Figure S3B ). Analogous to treatment with OTs, transfection with MYB siRNAs resulted in downregulation of CCNB1, CDK1, MCM4, and CD34 ( Figure 2F ). Our results indicate that these genes are MYB targets in T-ALL that are downregulated following OT treatment.

Both bardoxolone methyl and omaveloxolone inhibit NF-κB (40, 41), a signaling pathway that has been shown to positively regulate MYB transcription in murine erythroleukemia cells (42, 43). To test whether NF-κB signaling is involved in MYB gene regulation also in human T-ALL cells, we treated MOLT-4 cells with two IKK-1 and IKK-2 inhibitors and measured MYB mRNA levels by qPCR ( Supplementary Figure S4A ). Neither of these inhibitors decreased MYB gene expression, which may indicate that OTs downregulate MYB gene activity independently of NF-κB signaling in T-ALL cells. Similarly, treatment with the cell-penetrating NRF2-activating peptide KI696 (30, 44) did not affect MYB gene expression in T-ALL cells ( Supplementary Figure S4B ), suggesting that OTs target other pathways than NF-κB and NRF2 to regulate the MYB gene in T-ALL cells.

To study the effects of OT treatment on critical cellular processes in T-ALL cells, we performed RNA-seq analysis of MOLT-4 cells treated with 375 nM bardoxolone methyl or omaveloxolone for 24 h. Global gene expression analysis showed that OT treated cells clustered separately from control cells ( Figure 3A ). There was a significant overlap of differentially regulated genes between the two drugs ( Figures 3A–C ). To investigate if OT treatment affects the MYB-driven transcriptional program in T-ALL cells, we used GSEA to compare our RNA-seq data with a previously published dataset (45) of direct MYB target genes in hematopoietic progenitor cells ( Figures 3D, E ). OT treatment downregulated genes that are normally activated by MYB (e.g. MYC, TCF7, MAT2A) and upregulated known MYB-repressed genes (e.g. CEBPB, IL17RA, BCL6). Thus, our findings indicate that treatment with bardoxolone methyl and omaveloxolone can reverse the MYB-driven transcriptional program in T-ALL cells. GSEA also revealed activation of the p53 and apoptosis pathways in OT treated cells as well as decreased expression of genes regulating cell cycle progression, metabolism, and DNA repair ( Figures 4A, B ). Moreover, gene ontology analysis showed that genes upregulated by OT treatment are involved in cell differentiation and cell morphogenesis ( Figure 4C ) whereas genes downregulated are involved in ribosome biogenesis, RNA processing/metabolism, and nucleotide synthesis ( Figure 4D ). These results imply impaired synthesis of nucleic acids, RNA maturation, and translation in OT treated cells. Taken together, our RNA-seq data indicate that OT treatment reverses a MYB-driven gene expression program, inhibits cell proliferation, and induces differentiation and apoptosis in T-ALL cells.

To investigate if T-ALL cell viability is affected by treatment with OTs, we treated T-ALL cell lines derived from four cases with 50-1000 nM of bardoxolone methyl or omaveloxolone for 72 h ( Figure 5A ). The treatments resulted in a dose-dependent decrease in cell viability at nanomolar concentrations. There was also a significant induction of apoptosis in T-ALL cells after treatment with both drugs ( Figure 5B ). Importantly, HMC control cells were unaffected at low nanomolar concentrations and showed markedly less apoptosis at 500 nM compared with T-ALL cells. Cell cycle analysis showed that OT treated T-ALL cells displayed a marked G2/M arrest ( Figure 5C ). This was in agreement with both OT treatment and MYB knockdown causing a decreased expression of the G2 regulating genes CCNB1 and CDK1 ( Figures 2C, F ). OT treated cells also displayed a reduced entry into the S phase and an increased accumulation of apoptotic cells in sub G1 ( Figure 5C ). These results, which corroborate our findings from the global gene expression analysis ( Figures 3 , 4 ), show that T-ALL cells are sensitive to OT treatment.

Since genes involved in DNA repair were downregulated following OT treatment ( Figure 4B ), we tested whether the OTs could synergize with the DNA-damaging agent doxorubicin that is part of the standard therapy of T-ALL. Treatments with bardoxolone methyl or omaveloxolone combined with doxorubicin at IC₂₅ and IC₅₀ concentrations ( Figure 5A , Supplementary Figure S5 ) resulted in a significant synergistic decrease in cell viability at both concentrations in MOLT-4 cells ( Figure 6 ). These results suggest that synthetic OTs may sensitize T-ALL cells to chemotherapy.