MiR-15a-5p Confers Chemoresistance in Acute Myeloid Leukemia by Inhibiting Autophagy Induced by Daunorubicin

Anthracyclines remain a cornerstone of induction chemotherapy for acute myeloid leukemia (AML). Refractory or relapsed disease due to chemotherapy resistance is a major obstacle in AML management. MicroRNAs (miRNAs) have been observed to be involved in chemoresistance. We previously observed that miR-15a-5p was overexpressed in a subgroup of chemoresistant cytogenetically normal AML patients compared with chemosensitive patients treated with daunorubicin and cytarabine. MiR-15a-5p overexpression in AML cells reduced apoptosis induced by both drugs in vitro. This study aimed to elucidate the mechanisms by which miR-15a-5p contributes to daunorubicin resistance. We showed that daunorubicin induced autophagy in myeloid cell lines. The inhibition of autophagy reduced cell sensitivity to daunorubicin. The overexpression of miR-15a-5p decreased daunorubicin-induced autophagy. Conversely, the downregulation of miR-15a-5p increased daunorubicin-induced autophagy. We found that miR-15a-5p targeted four genes involved in autophagy, namely ATG9a, ATG14, GABARAPL1 and SMPD1. Daunorubicin increased the expression of these four genes, and miR-15a-5p counteracted this regulation. Inhibition experiments with the four target genes showed the functional effect of miR-15a-5p on autophagy. In summary, our results indicated that miR-15a-5p induces chemoresistance in AML cells through the abrogation of daunorubicin-induced autophagy, suggesting that miR-15a-5p could be a promising therapeutic target for chemoresistant AML patients.

Evi1 Counteracts Anti-Leukemic and Stem Cell Inhibitory Effects of All-Trans Retinoic Acid on Flt3-ITD/Npm1c-Driven Acute Myeloid Leukemia Cells

All-trans retinoic acid (atRA) has a dramatic impact on the survival of patients with acute promyelocytic leukemia, but its therapeutic value in other types of acute myeloid leukemia (AML) has so far remained unclear. Given that AML is a stem cell-driven disease, recent studies have addressed the effects of atRA on leukemic stem cells (LSCs). atRA promoted stemness of MLL-AF9-driven AML in an Evi1-dependent manner but had the opposite effect in Flt3-ITD/Nup98-Hoxd13-driven AML. Overexpression of the stem cell-associated transcription factor EVI1 predicts a poor prognosis in AML, and is observed in different genetic subtypes, including cytogenetically normal AML. Here, we therefore investigated the effects of Evi1 in a mouse model for cytogenetically normal AML, which rests on the combined activity of Flt3-ITD and Npm1c mutations. Experimental expression of Evi1 on this background strongly promoted disease aggressiveness. atRA inhibited leukemia cell viability and stem cell-related properties, and these effects were counteracted by overexpression of Evi1. These data further underscore the complexity of the responsiveness of AML LSCs to atRA and point out the need for additional investigations which may lay a foundation for a precision medicine-based use of retinoids in AML.

Genetic heterogeneity of cytogenetically normal AML with mutations of CEBPA

Biallelic mutations of the CCAAT/enhancer binding protein α (CEBPA) gene define a distinct genetic entity of acute myeloid leukemia (AML) with favorable prognosis. The presence of GATA2 and CSF3R mutations that are specifically associated with this subgroup but not mutated in all samples suggests a genetic heterogeneity of biCEBPA-mutated AML. We characterized the mutational landscape of CEBPA-mutated cytogenetically normal AML by targeted amplicon resequencing. We analyzed 48 biallelically mutated CEBPA (biCEBPA), 32 monoallelically mutated CEBPA (moCEBPA), and 287 wild-type CEBPA (wtCEBPA) patient samples from German AML Cooperative Group studies or registry. Targeted sequencing of 42 genes revealed that moCEBPA patients had significantly more additional mutations and additional mutated genes than biCEBPA patients. Within the group of biCEBPA patients, we identified 2 genetic subgroups defined by the presence or absence of mutations in chromatin/DNA modifiers (C), cohesin complex (C), and splicing (S) genes: biCEBPACCSpos (25/48 [52%]) and biCEBPACCSneg (23/48 [48%]). Equivalent subgroups were identified in 51 biCEBPA patients from the Cancer Genome Project. Patients in the biCEBPACCSpos group were significantly older and had poorer overall survival and lower complete remission rates following intensive chemotherapy regimens compared with patients in the biCEBPACCSneg group. Patients with available remission samples from the biCEBPACCSpos group cleared the biCEBPA mutations, but most had persisting CCS mutations in complete remission, suggesting the presence of a preleukemic clone. In conclusion, CCS mutations define a distinct biological subgroup of biCEBPA AML that might refine prognostic classification of AML. This trial was registered at www.clinicaltrials.gov as #NCT00266136 and NCT01382147.

Response and survival rates with frontline hypomethylating agent (HMAs) in favorable risk AML.

7039 Background: HMAs are an accepted frontline therapy for AML patients (pts) who are unfit for intensive induction therapy (IIT), particularly pts with unfavorable cytogenetics and/or p53 mutations. However, little is known about the response of favorable risk AML to HMAs. We previously reported that NPM1 mutated and/or CD34- AML status were predictors of response to HMAs. Here, we evaluated responses to frontline HMAs in AML. Methods: A total of 117 patients with de novo AML diagnosed between 7/2013 and 9/2016 were evaluated based on pt and disease related variables, overall response rate (ORR = CR + CR with incomplete count recovery + hematologic remission (ANC > 1000/µL, Hgb > 10g/dL, Plts > 100,000/µL, & no circulating blasts)), and overall survival (OS). Categorical variables were compared using Fisher’s exact test. Kaplan Meier methods estimated survival outcomes, and log rank tests compared survival between groups. Multivariable analyses were performed using Cox proportional hazards models. Results: 51 pts, considered unfit for IIT, received frontline HMAs. ORR and OS were highest in the ELN favorable risk AML pts (n = 13; ORR = 92%, p = .009; median OS = 17.5 months, p = .022). Among 41 NPM1 mutated pts, 15 received HMAs; and 26 received intensive induction. ORRs were 73% and 84%, respectively (p = .434). No difference was found in OS distributions between the HMA and IIT groups in univariate and multivariate (adjusted for age and FLT3 status) models (p = .329 and .241, respectively). Interestingly, ORR was 100% among 9 HMA-treated pts with NPM1 mutated, CD34-, FLT3/ITD-, cytogenetically normal AML. Conclusions: HMA therapy is highly effective frontline treatment in favorable risk AML pts considered unfit for IIT. Survival results with HMAs in NPM1 mutated AML are comparable to those of fitter pts treated with IIT. In selected favorable risk pts considered unfit for standard induction, HMAs can be a successful bridge to potentially curative therapy, including more intensive therapy or transplant. Cytogenetically normal AML with an isolated NPM1 mutation and CD34- status appears to be exceptionally responsive to frontline treatment with HMAs. Prospective validation of these findings is necessary.

Utility of Next Generation Sequencing in Prognostication and Therapeutic Decision Making in Cytogenetically Normal AML with DNMT3A Mutations

Background: The prognostication of cytogenetically normal AML (CN-AML) continues to evolve with the use of NGS-based risk stratification. Emerging data indicate that the presence of additional mutations in good- and intermediate-risk patients (as defined by conventional cytogenetic and molecular analyses) changes the behavior of their disease, suggesting that more personalized treatment approaches are needed. Methods: We analyzed the mutational profile of newly diagnosed AML patients with DNMT3A mutations (N = 48) seen at our center and described their clinical characteristics and associated mutations. These patients were identified as part of the Advanced Genomics in Leukemia (AGILE) clinical project currently underway at the Princess Margaret Cancer Centre, Toronto. NGS molecular profiling was performed using the TruSight Myeloid Sequencing Panel (TMSP; Illumina) on the MiSeq benchtop genome sequencer (Illumina). This process permitted profiling of 54 genes (39 in the hotspot region; 15 complete coding region coverage) using amplicon-based library preparation and sequencing by synthesis. 160 patients with newly diagnosed AML were evaluated for this project from February 2015 to February 2016. All were analyzed in parallel by the standard cytogenetic and molecular (NPM1, FLT3-ITD/TKD) diagnostic algorithm. Results: 48 unique patients were identified bearing mutated DNMT3A. Of these, 31 patients had a normal karyotype. Their clinical characteristics are depicted in Table 1. The R882X DNMT3A variant was detected in 12 patients. A total of 100 additional mutations were identified on sequencing (Figure1). The most common associated mutations were in NPM1 (38.7%) followed by IDH1 (29%), RUNX1 (25.8%) and TET2 (22.6%). PTPN11 mutations were identified in 6 patients, 75% of which also had mutated NPM1. Two patients were found to have biallelic CEBPA mutations. Potential gene-gene interactions were also examined and led to the identification of subgroups such as NPM1-FLT3-DNMT3A mutated (n=4), DNMT3A-IDH2R140 (n=3) and DNMT3A-IDH2R172 (n=3) based on recent data by Papaemmanuil, et al (New England Journal of Medicine, 2016) defining these subgroups as being prognostically relevant.( Patients >60 years of age had more frequent mutations in NRAS, BCOR, BCORL1 and TET2. NRAS and BCOR mutations were mutually exclusive with NPM1. RUNX1, IDH2, SRSF2 and U2AF1 mutations were also seen exclusively in the NPM1 negative group. Eligible patients (n=25) received induction therapy with daunorubicin and cytarabine leading to CR1 in 18 patients (72%). Primary induction failure occurred in 6 cases (24%). All 6 cases were NPM1 negative and had missense mutations in DNMT3A including 3 R882X variants. (Figure 2) Additional mutations were identified in IDH1, RUNX1 and/or TET2 in all 6 patients. During the median follow up of 8 months (range 1 - 15 months), 4 patients relapsed, 2 of which had mutated NPM1 with wild type FLT3-ITD. Sixteen patients are alive at this point and 12 are in CR, six having received allogeneic stem cell transplants. One patient with relapsed disease entered a clinical trial of an IDH1 inhibitor Conclusion: Genomic analysis is increasingly recognized as a vital adjunct to conventional diagnostic and prognostic approaches. With ongoing advancements in technology leading to increasing cost effectiveness and decreased turnaround times, the use of NGS is likely to become an up-front investigation resulting in a more personalized approach to therapy. Also, unique patient subgroups defined by gene-gene interactions can be identified to further predict clinical behavior and potentially identify druggable targets for therapy. Disclosures Gupta: Novartis: Consultancy, Honoraria, Research Funding; Incyte Corporation: Consultancy, Research Funding. Schimmer:Novartis: Honoraria. Yee:Novartis Canada: Membership on an entity's Board of Directors or advisory committees, Research Funding. Kamel-Reid:BMS: Research Funding. Schuh:Amgen: Membership on an entity's Board of Directors or advisory committees.