Preclinical efficacy of CDK7 inhibitor-based combinations against myeloproliferative neoplasms transformed to AML

Abstract

Rising blast percentage or secondary acute myeloid leukemia (sAML) transformation in myeloproliferative neoplasms (MPNs) leads to JAK1/2 inhibitor (JAKi) therapy resistance and poor survival. Here, we demonstrate that treatment with the CDK7 inhibitor (CDK7i) SY-5609 depletes phenotypically characterized post-MPN sAML stem/progenitor cells. In cultured post-MPN sAML SET2, HEL and patient-derived (PD) post-MPN sAML cells, SY-5609 treatment inhibited growth and induced lethality while sparing normal cells. RNA-sequencing analysis after SY-5609 treatment reduced mRNA expression of MYC, MYB, CDK4/6, PIM1, and CCND1 but increased expression of CDKN1A and BCL2L1. Mass spectrometry of SY-5609-treated MPN-sAML cells also reduced c-Myc, c-Myb, PIM1, and CDK4/6 but increased p21, caspase-9, and BAD protein levels. CRISPR-mediated CDK7 depletion also reduced cell viability of HEL cells. Cytometry by time of flight (CyTOF) analysis of SY-5609-treated PD post-MPN sAML stem/progenitor cells showed reduced c-Myc, CDK6, and PU.1 but increased protein levels of CD11b, p21, and cleaved caspase-3. Cotreatment with SY-5609 and ruxolitinib was synergistically lethal in HEL, SET2, and PD post-MPN sAML cells. A CRISPR screen in sAML cells revealed BRD4, CBP, and p300 as codependencies with CDK7i. Accordingly, cotreatment with SY-5609 and the bromodomain and extra-terminal protein inhibitor (BETi) OTX015 or pelabresib or the CBP/p300 inhibitor GNE-049 was synergistically lethal in MPN-sAML cells (including those exhibiting TP53 loss). Finally, in the HEL-Luc/GFP xenograft model, compared with each agent alone, cotreatment with SY-5609 and OTX015 reduced sAML burden and improved survival without host toxicity. These findings demonstrate promising preclinical activity of CDK7i-based combinations with BETi or CBP/p300 inhibitor against advanced MPNs, including post-MPN sAML.

Graphical abstract

Figure 1.

Treatment with CDK7i induced loss of viability and KO of CDK7 significantly attenuated cell viability and colony growth in post-MPN sAML cells. (A-B) SET2 and HEL92.1.7 cells were treated with the indicated concentrations of SY-1365 or SY-5609 for 96 hours. At the end of treatment, relative cell viability was determined using a CellTiter-Glo assay. The percentage of viable cells in each condition was normalized relative to untreated control cells. Columns represent the mean of 2 independent experiments performed in duplicate; bars, standard error of the mean (SEM). ∗∗∗∗P < .001, as determined by a 2-tailed, unpaired t test in GraphPad V10. (C-D) SET2 and HEL92.1.7 cells were treated with the indicated concentrations of SY-1365 or SY-5609 for 96 hours. After this, the cells were washed with 1× PBS and stained with TO-PRO-3 iodide. The percentage of TO-PRO-3 iodide–positive nonviable cells was determined by flow cytometry. Columns represent the mean of 3 experiments; bars, SEM. ∗∗∗P < .005; ∗∗∗∗P < .001, as determined by a 2-tailed, unpaired t test in GraphPad V10. (E) PD CD34⁺ sAML cells (n = 6) were treated with the indicated concentrations of SY-1365 or SY-5609 for 72 hours. Then, the cells were washed with 1× PBS and stained with TO-PRO-3 iodide. The percentage of TO-PRO-3 iodide–positive nonviable cells was determined by flow cytometry. Columns indicate the mean of 6 samples; bars, SEM. ∗P < .05; ∗∗P < .01; ∗∗∗P < .005, as determined by a 2-tailed, unpaired t test in GraphPad V10. (F) Normal CD34⁺ HPCs (from umbilical cord blood) were treated with the indicated concentrations of SY-1365 or SY-5609 for 72 hours. At the end of treatment, the cells were washed with 1× PBS and stained with TO-PRO-3 iodide. The percentage of TO-PRO-3 iodide–positive nonviable cells was determined by flow cytometry. Columns represent the mean of 4 samples; bars, SEM. (G) Immunoblot of HEL92.1.7 cells transfected with sgCtrl or sgCDK7 for 5 days. (H) Quantification of CDK7 KO by densitometry analysis. ∗∗∗∗P < .001, as determined by a 2-tailed, unpaired t test in GraphPad V10. (I) Cell cycle status of sgCtrl and sgCDK7 transfected cells 5 days after transfection as determined by flow cytometry. ∗P < .05; ∗∗P < .01, as determined by a 2-tailed, unpaired t test in GraphPad V10. (J-K) HEL92.1.7 cells transfected with sgCtrl or sgCDK7 were cultured for 6 days. Then the percentage cell viability was determined by CellTiter-Glo assay. Cells were also plated for colony growth assay in methylcellulose and incubated for 7 days. ∗∗∗P < .005, as determined by a 2-tailed, unpaired t test in GraphPad V10. (L) PD post-MPN sAML cells were transfected with sgCtrl or sgCDK7 were cultured for 6 days. Then the relative cell viability was determined by CellTiter-Glo assay. ∗∗∗P < .005, as determined by a 2-tailed, unpaired t test in GraphPad V10. HPC, hematopoietic progenitor cell; sgCtrl, single guide control.

Figure 2.

Treatment with SY-5609 caused negative enrichment of MYC targets, E2F targets, cell cycle checkpoints, and protein translation initiation and elongation in sAML cells. (A-B) HEL92.1.7 cells (biologic replicates) and PD JAK2-V617F–expressing sAML cells (biologic replicates) were treated with 100 nM of SY-5609 for 16 hours. Total RNA was isolated and used for RNA-seq analysis. Gene set enrichment analysis of SY-5609 treated cells compared with HALLMARK and REACTOME pathway data sets. All false discovery rate (FDR) q values are <0.1. (C-D) Enrichment plots of SY-5609–treated HEL92.1.7 and PD JAK2-V617F–expressing sAML cells compared with HALLMARK_MYC_TARGETS_V1 and V2 data sets. (E) Volcano plot of significantly altered mRNAs (P < .05) in SY-5609–treated HEL92.1.7 cells compared with untreated control cells. (F) Volcano plot of significantly altered mRNAs (P < .05) in SY-5609–treated PD JAK2-V617F–expressing sAML cells compared with untreated control cells. NES, normalized enrichment score.

Figure 3.

Treatment with SY-5609–depleted expression of MYC and E2F target gene sets in HEL cells and JAK-STAT3 signaling proteins and E2F target gene sets in PD mtCALR sAML cells as well as depleted protein expression of c-Myc, CDK6, and PU.1 in phenotypically defined sAML stem/progenitor cells. (A-B) HEL92.1.7 cells were treated with 100 nM of SY-5609 for 24 hours in biologic duplicates. At the end of treatment, cells were harvested and used for whole proteome tandem mass spectrometry. Gene set enrichment analysis was conducted with the protein expression signature of SY-5609–treated cells vs control cells against HALLMARK MYC_TARGETS and E2F_TARGETS data sets. All q values are <0.1. (C) PD mtCALR sAML cells were treated with 100 nM of SY-5609 for 24 hours in biologic duplicates. At the end of treatment, cells were harvested and used for whole proteome tandem mass tag (tmt) mass spectrometry. Gene set enrichment analysis was conducted with the SY-5609–mediated protein expression signature against HALLMARK data sets. All q values are <0.1. (D) Volcano plot of protein expression alterations in SY-5609–treated HEL92.1.7 cells compared with control cells. The threshold is proteins with greater than a 1.2-fold change up or down and a P value <.05. (E-F) PD CD34⁺ sAML cells (n = 3) were treated with 100 nM of SY-5609 for 24 hours. Cells were used for CyTOF analysis with a cocktail of rare metal–tagged antibodies to define stem/progenitor cells and other sAML-relevant oncoproteins. The heat map (E) shows the log2 fold change (100 nM of SY-5609 for 24 hours compared with untreated control cells) determined by CyTOF analyses in the 3 PD sAML samples. (F) The graph shows the percentage of PD CD34⁺ sAML cells expressing a stem/progenitor cell phenotype (CLEC12A hi, CD123 hi, CD99 hi, CD33 hi, and CD11b lo) in the untreated control cells and after SY-5609 treatment. Hi, high; lo, low; mtCALR, mutant calreticulin; NES, normalized enrichment score.

Figure 4.

A protein domain–specific CRISPR screen identified druggable codependencies with SY-5609 treatment in HEL92.1.7 and SET2 cells, and cotreatment with SY-5609 and ruxolitinib, BETi, or HAT inhibitor induced synergistic lethality in post-MPN sAML cells. (A) HEL92.1.7 and SET2 CRISPR-associated protein 9 (Cas9)-expressing cells were transduced with a library of sgRNAs against epigenetic modifier proteins (1390 sgRNAs including 50 negative control sgRNAs). Eight days after transduction, cells were treated with 100 nM of SY-5609 for 4 days. Viable cells were harvested, genomic DNA was isolated, and polymerase chain reaction amplified with primers surrounding the sgRNAs. Amplicon-Seq (next-generation sequencing [NGS]) was performed to analyze sgRNA abundance in the control and SY-5609–treated cells. Panels show the log2 fold change in sgRNAs that were further depleted by treatment with SY-5609 compared with the day 12 untreated controls (P < .05). (B) HEL92.1.7, SET2, and PD sAML cells were treated with SY-5609 and/or ruxolitinib, OTX015, CPI-0610 (pelabresib), or GNE-049 for 72 hours. At the end of treatment, cells were washed with 1× PBS and stained with TO-PRO-3 iodide. The percentage of nonviable cells were determined by flow cytometry. Delta Synergy scores were determined with the SynergyFinder web application using the zero interaction potency (ZIP) method. Scores >1.0 indicate a synergistic interaction between the 2 agents in the combination. N.D., not done.

Figure 5.

Treatment with CDK7i induced loss of viability in JAKi-T/R cells, and pretreatment with SY-5609 partially reversed resistance to ruxolitinib or BETi. (A-B) SET2 JAKi-T/R and HEL JAKi-T/R cells were treated with the indicated concentrations of SY-5609 for 24 hours. After this, cells were fixed with 70% ethanol; then washed with 1× PBS and stained with propidium iodide. The cell cycle status was determined by flow cytometry. ∗P < .05; ∗∗∗P < .005, compared with untreated control cells (determined by a 2-tailed, unpaired t test in GraphPad V10). (C) SET2 JAKi-T/R and HEL JAKi-T/R cells were treated with the indicated concentrations of SY-5609 for 72 hours. After this, the cells were washed with 1× PBS and stained with TO-PRO-3 iodide. The percentage of TO-PRO-3 iodide–positive nonviable cells was determined by flow cytometry. Columns represent the mean of 3 experiments; bars, SEM. (D-E) SET2 JAKi-T/R and HEL JAKi-T/R cells were pretreated with 100 nM of SY-5609 for 24 hours. Then, cells were treated with the indicated concentrations of ruxolitinib for 48 hours, and the percentage of annexin V–positive apoptotic cells was determined by flow cytometry. ∗∗P < .01; ∗∗∗P < .005; ∗∗∗∗P < .001, compared with SY-5609–treated cells not treated with ruxolitinib (determined by a 2-tailed, unpaired t test in GraphPad V10). (F) Immunoblot analysis of cell lysates from SET2, SET2-JAKi–T/R, and SET2-OTX–T/R cells treated with the indicated concentration of SY-5609 for 24 hours. The expression levels of GAPDH in the cell lysates served as the loading control. Numbers beneath the bands represent densitometry analysis compared with SET2 parental cells. (G) SET2 and SET2-OTX–T/R cells were treated with the indicated concentrations of SY-5609 for 72 hours. After this, the cells were washed with 1× PBS and stained with TO-PRO-3 iodide. The percentage of TO-PRO-3 iodide–positive nonviable cells was determined by flow cytometry. Columns represent the mean of 3 experiments; bars, SEM. (H) SET2-OTX–T/R cells were pretreated with 100 nM of SY-5609 for 24 hours. Then, cells were treated with the indicated concentrations of OTX015 for 72 hours, and the percentage of TO-PRO-3 iodide–positive nonviable cells were determined by flow cytometry. ∗P < .05; ∗∗∗P < .005, compared with SY-5609–treated cells with no OTX015 treatment (determined by a 2-tailed, unpaired t test in GraphPad V10). Pre-Rx, pretreatment.

Figure 6.

Treatment with SY-5609 dose-dependently induced G1 accumulation, reduced the percentage of murine JAK2-V617F, TP53-KO cells in S phase, and significantly reduced leukemia burden in a murine JAK2-V617F, TP53-KO erythroleukemia transplant model. (A) Murine JAK2-V617F, TP53-KO cells were treated with the indicated concentrations of SY-5609 for 24 hours. After this, cells were fixed with 70% ethanol; then washed with 1× PBS and stained with propidium iodide. The cell cycle status was determined by flow cytometry. ∗P < .05; ∗∗P < .01; ∗∗∗P < .005, compared with untreated control cells (determined by a 2-tailed, unpaired t test in GraphPad V10). (B) Murine JAK2-V617F, TP53-KO, and JAK2 WT, TP53 WT cells were treated with the indicated concentrations of SY-5609 for 72 hours. At the end of treatment, relative cell viability was determined using a CellTiter-Glo assay. The percentage of viable cells in each condition was normalized relative to the untreated control cells. Columns represent the mean of 2 independent experiments performed in duplicate; bars, SEM. (C) Murine JAK2-V617F, TP53-KO erythroleukemia cells were treated with SY-5609 and/or ruxolitinib or OTX015 for 72 hours. Relative cell viability was determined using a CellTiter-Glo assay. The percentage of nonviable cells in each condition was normalized relative to the untreated control cells. Delta synergy scores were determined with the SynergyFinder web application using the ZIP method. Scores >1.0 indicate a synergistic interaction between the 2 agents. (D) Murine JAK2-V617F, TP53-KO erythroleukemia cells (1000 cells) were combined with 1 million total bone marrow cells from JAXBOY mice and infused by lateral tail vein into lethally irradiated (6 Gy) JAXBOY mice. After engraftment, mice were treated with vehicle or SY-5609 (1.5 mg/kg B.I.D. × 5 days per week) for 3 weeks. Mice were euthanized, and the percentage of dTomato-positive cells in the bone marrow, spleen, and liver was determined by flow cytometry. ∗∗∗P < .005, compared with vehicle-treated mice (determined by a 2-tailed, unpaired t test in GraphPad V10). B.I.D., twice daily; Rx, treatment; WT, wild-type.

Figure 7.

Treatment with SY-5609–based combinations reduced leukemia burden and significantly improved survival of NSG mice bearing HEL92.1.7 xenografts. (A) Schematic of in vivo treatment schedules in NSG mice. (B) Total photon counts (flux; determined by bioluminescent imaging) in NSG mice (n = 10) engrafted with luciferized HEL92.1.7 cells and treated with the indicated dose of SY-5609 and/or ruxolitinib for 3 weeks. Significance between cohorts was determined by a 2-tailed, unpaired t test in GraphPad V10. ∗P < .05; ∗∗∗P < .005. (C) Representative bioluminescent images of mice from panel B. (D) Kaplan-Meier survival plot of NSG mice treated with SY-5609 alone, ruxolitinib alone, or SY-5609 and ruxolitinib for 4 weeks. Significance was calculated by a Mantel-Cox log-rank test. ∗∗P < .01; ∗∗∗P < .005; ∗∗∗∗P < .001. (E) Total photon counts (flux; determined by bioluminescent imaging) in NSG mice (n = 10) engrafted with luciferized HEL92.1.7 cells and treated for 3 weeks with SY-5609 and/or OTX015 at the indicated doses. Significance between cohorts was determined by a 2-tailed, unpaired t test in GraphPad V10; ∗P < .05; ∗∗P < .01. (F) Representative bioluminescent images of mice from panel E. (G) Kaplan-Meier survival plot of NSG mice engrafted with luciferized HEL92.1.7 cells and treated with SY-5609 and/or OTX015 at the indicated doses for 4 weeks. Significance between cohorts was determined by a Mantel-Cox log-rank test. ∗∗∗P < .005; ∗∗∗∗P < .001. Rx, treatment.