Combination of azacitidine and enasidenib enhances leukemic cell differentiation and cooperatively hypomethylates DNA

•Enasidenib is an inhibitor of mutant IDH2 that releases the block in cellular differentiation imposed by aberrant production 2-hydroxyglutarate.•Mechanisms action azacitidine and enasidenib involve regulation epigenome.•The enhanced anti-leukemic activity combining leads to greater reductions DNA methylation. Azacitidine are two therapies available for treatment acute myelogenous leukemia (AML), mechanisms these drugs alteration We hypothesized a combination agents could have interactive effects on methylation enhance mIDH2 cells. Combination TF-1 cells overexpressing IDJ2R140Q through increased hemoglobinization hemoglobin ? RNA expression compared with single agents. Furthermore, primary AML samples (IDH2R140Q or R172K), reduced CD34+ CD15+ extent than attained To explore mechanism treatment, epigenome was analyzed profiling 5-hydroxymethylcytosine (5hmC) 5-methylcytosine (5mC) changes. Enasidenib alone 5hmC, consistent reactivation ten-eleven-translocation (TET) enzyme activity. Compared alone, 5mC levels at numbers sites loci were significantly enriched regions 5hMC (25.8% vs. 7.4%). Results model which enasidenib-mediated enzymes cooperates azacitidine-mediated inhibition methyltransferase enzymes, leading erythroid differentiation. The metabolic isocitrate dehydrogenase 2 (IDH2) mutated hallmark [1Figueroa ME Lugthart S Li Y et al.DNA signatures identify biologically distinct subtypes myeloid leukemia.Cancer Cell. 2010; 17: 13-27Abstract Full Text PDF PubMed Scopus (663) Google Scholar,2Bullinger L Ehrich M Dohner K al.Quantitative predicts survival adult leukemia.Blood. 115: 636-642Crossref (130) Scholar]. regulated addition methyl groups cytosines methyltransferases (DNMTs) produce [3Jones PA Liang G. Rethinking how patterns maintained.Nat Rev Genet. 2009; 10: 805-811Crossref (559) Scholar] removal including oxidation followed further conversion unmethylated cytosine [4Kohli R Zhang Y. TET TDG dynamics demethylation.Nature. 2013; 502: 472-479Crossref (1020) In hematopoietic express IDH1/2, high 2-hydroxyglutarate (2-HG) inhibit [5Xu W Yang H Liu al.Oncometabolite competitive alpha-ketoglutarate-dependent dioxygenases.Cancer 2011; 19: 17-30Abstract (1912) Scholar], impaired [6Figueroa Abdel-Wahab O Lu C al.Leukemic IDH1 mutations result hypermethylation phenotype, disrupt TET2 function, impair differentiation.Cancer 18: 553-567Abstract (1966) Scholar, 7Kernytsky A Wang F Hansen E al.IDH2 mutation-induced histone progressively reversed small-molecule inhibition.Blood. 2015; 125: 296-303Crossref (129) 8Wang Travins J DeLaBarre B al.Targeted induces differentiation.Science. 340: 622-626Crossref (613) Scholar].Enasidenib used relapsed/refractory harboring [9IDHIFA® (enasidenib) prescribing information. Celgene Corp., Summit, NJAugust 2017Google DNMT broadly [10Vidaza® (azacitidine) NJSeptember 2018Google Scholar,11Vidaza® summary product characteristics. Europe Limited, Uxbridge, UK2015Google As both regulate methylation, we combined epigenetic effect occur DNMTs.MethodsMaterials methods described detail Supplementary Data (online only, www.exphem.org).Results discussionCombination enhances epoetin-induced IDH2-R140Q cellsTo examine differentiation, stably expressing variant evaluated assay. These micromolar 2-HG, characterized immature inability differentiate response erythropoietin (EPO) [7Kernytsky Scholar,8Wang Cells treated sequentially, preceding (Figure 1A). Erythroid visualization cell color, as indicator hemoglobinization, quantification (HBG) expression.Single-agent induced dose-dependent manner, did single-agent lesser 1B). Treatment agents, greatest observed highest dose enasidenib. Enhanced also using concurrent dosing schedule; however, death (1 µmol/L) this schedule (Supplementary Figure E1, online www.exphem.org). Drug treatments sequential had little growth up 4 days dimethyl sulfoxide (DMSO) control (data not shown). By contrast, use schedule, decreased 30% DMSO control, results.Single-agent (0.2 1 ?mol/L) HBG 64- 176-fold, respectively, (0.3 24- 59-fold, vehicle 1C). µmol/L), 2.2- 2.8-fold, agent. Similarly, 5.7- 8.4-fold, agent.To cultures, bone marrow mononuclear obtained from patients IDH2R172 IDH2R140 treated, flow cytometry quantify populations CD34 marker stem/progenitor CD15 granulocytic and/or monocytic maturation. samples, either generally similar those whereas increasing doses inhibited IDH2WT drug 50% control. percentage (–30% –97% alone) (1.7- 11.6-fold) Figures E2 E3, www.exphem.org), indicating more effective inducing granulocytic/monocytic maturation IDH2-mutated samples.Synergy analyses performed multiple (BLISS [breaks labeling situ sequencing] BRAID [bivariate additive interacting doses] methods) endpoints measured quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) cytometry. indices calculated according BLISS independent multiplicative each outlined Table E1 summary, synergy method analysis KLF qRT-PCR, CD34+/CD38+ cytometry, CD34+/CD38? schedules. revealed significant predictions based positive ? values small confidence intervals qRT-PCR (concurrent schedule) (sequential schedule).Supplemental E1Synergy BRAID) qPCR cytometry.Interaction index (Bliss)BraidAssayMarkerTreatmentMinMaxSynergyKappa (95% CI)SynergyqPCRKLFSequential?1.130.44+0.19 (-0.44, 0.85)?HBGSequential?5.33?1.20??0.08 (-0.32, 0.29)?KLFConcurrent?1.02?0.25+9.28 (1.49, 42.54)+†Large CI interval Kappa might indicate violation assumptionHBGConcurrent?1.40?0.03?2.47 (0.16, 4.77)+Flow cytometryCD34+/CD38+Sequential?0.150.30+22.9 (3.62, 76.65)+*AZA 1µM has higher population AZA 0.3µM, fitting may fail (also noted large value)†Large assumptionCD34+/CD38-Sequential?0.060.30+4.06 (1.09, 8.16)+CD34+/CD38+Concurrent?0.220.20+6.08 (-.47, 49.67)?*AZA assumptionCD34+/CD38-Concurrent?0.020.11+32.41 (0.7, 100)+*AZA assumption value)† Large Open table new tab Epigenomic azacitidine–enasidenib molecular underlying interaction explored 5hmC changes (azacitidine, µmol/L; enasidenib, µmol/L).Effects hydroxymethylated immunoprecipation (hMeDIP) sequencing day 11, prior differentiating stimulus EPO (pre-EPO) 2A). predominantly (94,942 94,986 DhMRs differentially [DhMRs] 5hmC,), (78,625 83,065 5hmC,). increase many regions. its known action.Figure 2Epigenomic azacytidine (A) ?mol/L), ENA 11 days, hydroxymethylation hME-DIP sequencing. Each condition triplicate. DhMR reported comparison (B) without 1A. (C) proportions DMSs within (%) plotted 18 demethylated alone. Loci equal between included alone.View Image ViewerDownload Hi-res image Download (PPT)Supplemental E1Concurrent A) Concurrent (AZA, 0.3 ?M) (ENA, 0.2 cells; additional exposure throughout time course resulted cytotoxicity . B) pellets post-treatment, red coloration indicative hemoglobinization. C) P-values One-way ANOVA Dunnett's test vs DMSO/EPO *p ? 0.05. **p 0.01.View E2Combination A-B) BMMCs cultured vitro presence (AZA) (ENA) identified percentages CD45+ Percentage (left) fold change (right) (of cells) quantified normalized vehicle-treated after 7 treatment.View E3Flow gating strategy patient samples. representative shown Note lymphoid excluded gate. population) samples.View (PPT)Effects ERRBS (post–EPO-induced differentiation), methylated (DMSs) 2B). (77,947 65,285, respectively), azacitidine. At (81,278 77,947, respectively); 18, plus (141,096 respectively) E2, Single-agent minimal points.Supplemental E2Numbers (Days 18).Numbers Differentially Methylated LociDrug TreatmentDirectionDay (Pre-EPO)Day (Post-EPO-induced differentiation)AZAHypermethylation103294AZAHypomethylation7794765285ENAHypermethylation277ENAHypomethylation5626AZA+ENAHypermethylation159345AZA+ ENAHypomethylation81278141096 would versus our hypothesis, relationship drug-induced (71,953 loci). DMS categories: (1) (2) same This methodology because differences hMeDIP assays do permit direct individual (see details Data). proportion (13.6% 12.3%) 2C). results DNMTs, specific loci. increases azacitidine-mediated, accelerated cytosine.Despite several recent approvals AML, rational combinations can achieve frequent, deep, lasting responses still desperately needed. end, found relieves blockade suggest effect. evidenced HBF expression, reduction model, mutations, mature granulocyte/monocyte population. Kinetic evaluation only magnitude but rapid onset agent samples; no wild-type (Supplemental E4, Supplemental E5 E5, Considering rates well fact best clinical take cycles potential therapy implications.Supplemental E4Time Primary combination. 12 culture frequencies determined cells.View E5Time (PPT)The randomized phase portion study NCT02677922 evaluating safety efficacy injectable newly diagnosed IDH2. Preliminary data (overall 68% 42%, p = 0.0015; complete remission 12%, 0.0002) [12DiNardo CD Schuh AC Stein EM al.Enasidenib Plus improves overall (AML) mutations: interim II ongoing, study.Blood. 2019; 134 (643–643)PubMed hypothesis inhibiting “writer” (DNMT) while reactivating “eraser” lead profound reversal epigenomic marks enforce exact nature gene level yet been fully elucidated. DNMTs. MethodsMaterials Materials ob