MLL Fusion Proteins Research Articles

Abstract SY24-03: Chromatin regulation in cancer

Abstract Recent exom sequencing studies have discovered that a wide variety of human malignancies appear to have driving mutations in genes encoding subunits of chromatin regulatory complexes. In particular mutations in the subunits of the BAF (mSWI/SNF) complex (Khavari, P. et al. Nature 1993; Wu, J.I. et al. Cell 2009) are most common and found in over 20% of all human malignancies (Kadoch, C et al. Nature Genetics 2013). Mutations in subunits of the polycomb repressive complex 2 (PRC2) are also frequent and are found most commonly in the catalytic EZH2 subunit (Garroway and Lander Cell 2013; Dawson MA, and Kouzarides T. Cell. 2012). These two activities oppose one another and appear to form a dynamic balance that controls expression of genes involved in cell proliferation, genome protection and development. BAF complexes can function as either onocgenes or as tumor suppressors (Kadoch, C. and Crabtree Cell 2013; Kadoch et al Nature Genetics 2013, Dunaief JL et al Cell 1994). (see also the Abstract by Cigall Kadoch). In general these subunits behave as tumor suppressors. We find that BAF complexes are necessary for the binding and function of TopoIIa over the genome (Dykhuzien et al Nature 2013). Deletion of subunits such as Brg, BAF250a, BAF47 or BAF57 leads to the presence of anaphase bridges reflecting a failure of decatenation during mitosis. Deletion of several BAF subunits also evokes the decatenation checkpoint with cell cycle arrest that can be specifically repressed with the TopoIIa S1524A mutation that is unable to initiate the decatenation checkpoint. TopoIIa directly interacts with BAF250a through a region that contains several dozen mutations in human cancer. A small molecule screen for inhibitors that block the ability of BAF complexes to suppress the PRC1 subunit, BMI revealed several TopoIIa inhibitors (Dykhuzien et al Nature 2013). Surprisingly these TopoIIa inhibitors mimic both the repression of genes such as Bmi1, Ring1a and other polycomb subunits and also the activation of genes such as FGF4. These studies suggest that TopoIIa and BAF function together to both control transcription and also to resolve decatenated DNA during mitosis. The later may contribute to the role of BAF complexes as tumor suppressors. Interestingly, the most frequently mutated subunits in human disease, BAF250a, SS18, Bcl7 and others are not required for in vitro chromatin remodeling, ATP activity or complex assembly and have no yeast homologues. In addition, we find that deletion of subunits of the complexes do not change nucleosome positioning over the mammalian genome. These unexpected findings underline the importance of developing new techniques to study chromatin regulation. With this goal in mind we have developed strains of mice that allow one to direct a specific chromatin regulatory activity to one allele of a marked gene using small molecule inducers of proximity. Using this approach we have explored the kinetics and stability of heterchromatin formation (Hathaway, Bell et al Cell 2012; Hodges et al PNAS 2013) revealing that H3K9Me3 induced heterochromatin is limited by intrinsic reaction rates rather than borders (Hodges et al PNAS 2012). Additional studies directed at the mechanism of action of the oncogenic MLL-fusion proteins and the ATP-dependent remodeling complexes will be discussed. Citation Format: Gerald Crabtree. Chromatin regulation in cancer. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr SY24-03. doi:10.1158/1538-7445.AM2014-SY24-03

Tumor promoting role of the DNA damage response

Tumor promoting role of the DNA damage response

Importance of a specific amino acid pairing for murine MLL leukemias driven by MLLT1/3 or AFF1/4

Acute leukemias caused by translocations of the MLL gene at chromosome 11 band q23 (11q23) are characterized by a unique gene expression profile. More recently, data from several laboratories indicate that the most commonly encountered MLL fusion proteins, MLLT1, MLLT3, and AFF1 are found within a molecular complex that facilitates the elongation phase of mRNA transcription. Mutational analyses suggest that interaction between the MLLT1/3 proteins and AFF family proteins are required for experimental transformation of hematopoietic progenitor cells (HPCs). Here, we define a specific pairing of two amino acids that creates a salt bridge between MLLT1/3 and AFF proteins that is critically important for MLL-mediated transformation of HPCs. Our findings, coupled with the newly defined structure of MLLT3 in complex with AFF1, should facilitate the development of small molecules that block this amino acid interaction and interfere with the activity of the most common MLL oncoproteins.

Comparative functional analysis of the molecular network of 7 selected MLL fusion proteins

The Mixed Lineage Leukemia gene (MLL) is a frequent target of chromosomal rearrangements in human malignancies. MLL-rearrangements are primarily associated with pediatric leukemias with poor prognosis. Balanced translocations result in the fusion of the MLL gene to over 60 fusion partner genes, leading to the production of novel chimeric proteins. Even though critical effectors of distinct MLL fusion proteins have been identified, it is not clear if these effectors are conserved among all MLL fusion proteins or if different molecular mechanisms of transformation exist.

Psip1/Ledgf p75 restrains Hox gene expression by recruiting both trithorax and polycomb group proteins.

Trithorax and polycomb group proteins are generally thought to antagonize one another. The trithorax family member MLL (myeloid/lymphoid or mixed-lineage leukemia) is presumed to activate Hox expression, counteracting polycomb-mediated repression. PC4 and SF2 interacting protein 1 (PSIP1)/p75, also known as LEDGF, whose PWWP domain binds to H3K36me3, interacts with MLL and tethers MLL fusion proteins to HOXA9 in leukaemias. Here we show, unexpectedly, that Psip1/p75 regulates homeotic genes by recruiting not only MLL complexes, but also the polycomb group protein Bmi1. In Psip1−/− cells binding of Mll1/2, Bmi1 and the co-repressor Ctbp1 at Hox loci are all abrogated and Hoxa and Hoxd mRNA expression increased. Our data not only reveal a potential mechanism of action for Psip1 in the regulation of Hox genes but also suggest an unexpected interplay between proteins usually considered as transcriptional activators and repressors.

Functional characterisation of different MLL fusion proteins by using inducible Sleeping Beauty vectors

Our focus is the identification, characterisation and functional analysis of different MLL fusions. In general, MLL fusion proteins are encoded by large cDNA cassettes that are difficult to transduce into haematopoietic stem cells. This is due to the size limitations of the packaging process of those vector-encoded RNAs into retro- or lentiviral particles. Here, we present our efforts in establishing a universal vector system to analyse different MLL fusions. The universal cloning system was embedded into the backbone of the Sleeping Beauty transposable element. This transposon has no size limitation and displays no integration preference, thereby avoiding the integration into active genes or their promoter regions. We utilised this novel system to test different MLL fusion alleles (MLL-NEBL, NEBL-MLL, MLL-LASP1, LASP1-MLL, MLL-MAML2, MAML2-MLL, MLL-SMAP1 and SMAP1-MLL) in appropriate cell lines. Stable cell lines were analysed for their growth behaviour, focus formation and colony formation capacity and ectopic Hoxa gene transcription. Our results show that only 1/4 tested direct MLL fusions, but 3/4 tested reciprocal MLL fusions exhibit oncogenic functions. From these pilot experiments, we conclude that a systematic analysis of more MLL fusions will result in a more differentiated picture about the oncogenic capacity of distinct MLL fusions.

Proteasome Inhibitors Evoke Latent Tumor Suppression Programs in Pro-B MLL Leukemias through MLL-AF4

Chromosomal translocations disrupting MLL generate MLL-fusion proteins that induce aggressive leukemias. Unexpectedly, MLL-fusion proteins are rarely observed at high levels, suggesting excessive MLL-fusions may be incompatible with a malignant phenotype. Here, we used clinical proteasome inhibitors, bortezomib and carfilzomib, to reduce the turnover of endogenous MLL-fusions and discovered that accumulated MLL-fusions induce latent, context-dependent tumor suppression programs. Specifically, in MLL pro-B lymphoid, but not myeloid, leukemias, proteasome inhibition triggers apoptosis and cell cycle arrest involving activation cleavage of BID by caspase-8 and upregulation of p27, respectively. Furthermore, proteasome inhibition conferred preliminary benefit to patients with MLL-AF4 leukemia. Hence, feasible strategies to treat cancer-type and oncogene-specific cancers can be improvised through harnessing inherent tumor suppression properties of individual oncogenic fusions.

Downregulation of RUNX1/CBFβ by MLL fusion proteins enhances hematopoietic stem cell self-renewal

RUNX1/CBFβ (core binding factor [CBF]) is a heterodimeric transcription factor complex that is frequently involved in chromosomal translocations, point mutations, or deletions in acute leukemia. The mixed lineage leukemia (MLL) gene is also frequently involved in chromosomal translocations or partial tandem duplication in acute leukemia. The MLL protein interacts with RUNX1 and prevents RUNX1 from ubiquitin-mediated degradation. RUNX1/CBFβ recruits MLL to regulate downstream target genes. However, the functional consequence of MLL fusions on RUNX1/CBFβ activity has not been fully understood. In this report, we show that MLL fusion proteins and the N-terminal MLL portion of MLL fusions downregulate RUNX1 and CBFβ protein expression via the MLL CXXC domain and flanking regions. We confirmed this finding in Mll-Af9 knock-in mice and human M4/M5 acute myeloid leukemia (AML) cell lines, with or without MLL translocations, showing that MLL translocations cause a hypomorph phenotype of RUNX1/CBFβ. Overexpression of RUNX1 inhibits the development of AML in Mll-Af9 knock-in mice; conversely, further reducing Runx1/Cbfβ levels accelerates MLL-AF9-mediated AML in bone marrow transplantation assays. These data reveal a newly defined negative regulation of RUNX1/CBFβ by MLL fusion proteins and suggest that targeting RUNX1/CBFβ levels may be a potential therapy for MLLs.

Challenges and Opportunities in Targeting the Menin–MLL Interaction

Menin is an essential co-factor of oncogenic MLL fusion proteins and the menin-MLL interaction is critical for development of acute leukemia in vivo. Targeting the menin-MLL interaction with small molecules represents an attractive strategy to develop new anticancer agents. Recent developments, including determination of menin crystal structure and development of potent small molecule and peptidomimetic inhibitors, demonstrate the feasibility of targeting the menin-MLL interaction. On the other hand, biochemical and structural studies revealed that MLL binds to menin in a complex bivalent mode engaging two MLL motifs, and therefore inhibition of this protein-protein interaction represents a challenge. This review summarizes the most recent achievements in targeting the menin-MLL interaction as well as discusses potential benefits of blocking menin in cancer.

A distinct set of long non-coding RNAs in childhood MLL-rearranged acute lymphoblastic leukemia: biology and epigenetic target

Long non-coding RNAs (lncRNAs) have been recently found to be pervasively transcribed in human genome and link to diverse human diseases. However, the expression patterns and regulatory roles of lncRNAs in hematopoietic malignancies have not been reported. Here, we carried out a genome-wide lncRNA expression study in MLL-rearranged acute lymphoblastic leukemia (MLL-r ALL) and established lncRNA/messenger RNA coexpression networks to gain insight into the biological roles of these dysregulated lncRNAs. We detected a number of lncRNAs that were differentially expressed in MLL-r ALL samples compared with MLL-r wild-type and identified unique lncRNA expression patterns between MLL-r subtypes with different translocations as well as between infant MLL-r ALL with other MLL-r ALL patients, suggesting that they might be served as novel biomarkers for the disease. Importantly, several lncRNAs that correspond with membrane protein genes, including a lysosome-associated membrane protein, were identified. No such link between the membrane proteins and MLL-r leukemia has been reported previously. Impressively, the functional analysis showed that several lncRNAs corresponded to the expression of MLL-fusion protein target genes, including HOXA9, MEIS1, etc., while some other associated with histone-related functions or membrane proteins. Further experiments characterize the effect of some lncRNAs on MLL-r leukemia apoptosis and proliferation as the function of the coexpressed HOXA gene cluster. Finally, a set of lncRNAs epigenetically regulated by H3K79 methylation were also discovered. These findings may provide novel insights into the mechanisms of lncRNAs involved in the initiation of MLL-r leukemia. This is the first study linking lncRNAs to leukemogenesis.

MLL fusion proteins link transcriptional coactivators to previously active CpG-rich promoters

Mixed-lineage leukemia (MLL) maintains the expression of cellular memory genes during development, while leukemic MLL fusion proteins aberrantly maintain expression of hematopoietic stem cell program genes such as HOXA9 to cause leukemia. However, the molecular mechanism of gene activation is unclear. Here we show that only two functional modules are necessary and sufficient for target recognition: those that bind to non-methylated CpGs and di-/tri-methylated histone H3 lysine 36 (H3K36me2/3). An artificial protein composed of the two targeting modules and an interaction domain for AF4-family coactivators can functionally substitute for MLL fusion proteins. Because H3K36me2/3 markers are indicative of active transcription, MLL fusion proteins target previously active CpG-rich genes and activate transcription by recruiting coactivators thereto. Our results indicate that such chromatin context-dependent gene activation is the fundamental mechanism by which MLL fusion proteins maintain the expression of the cellular memory/hematopoietic stem cell program genes.

Classifying leukemia types with chromatin conformation data.

BackgroundAlthough genetic or epigenetic alterations have been shown to affect the three-dimensional organization of genomes, the utility of chromatin conformation in the classification of human disease has never been addressed.ResultsHere, we explore whether chromatin conformation can be used to classify human leukemia. We map the conformation of the HOXA gene cluster in a panel of cell lines with 5C chromosome conformation capture technology, and use the data to train and test a support vector machine classifier named 3D-SP. We show that 3D-SP is able to accurately distinguish leukemias expressing MLL-fusion proteins from those expressing only wild-type MLL, and that it can also classify leukemia subtypes according to MLL fusion partner, based solely on 5C data.ConclusionsOur study provides the first proof-of-principle demonstration that chromatin conformation contains the information value necessary for classification of leukemia subtypes.

PU.1 is essential for MLL leukemia partially via crosstalk with the MEIS/HOX pathway

Mixed lineage leukemia (MLL) fusion proteins directly activate the expression of key downstream genes such as MEIS1, HOXA9 to drive an aggressive form of human leukemia. However, it is still poorly understood what additional transcriptional regulators, independent of the MLL fusion pathway, contribute to the development of MLL leukemia. Here we show that the transcription factor PU.1 is essential for MLL leukemia and is required for the growth of MLL leukemic cells via the promotion of cell-cycle progression and inhibition of apoptosis. Importantly, PU.1 expression is not under the control of MLL fusion proteins. We further identified a PU.1-governed 15-gene signature, which contains key regulators in the MEIS-HOX program (MEIS1, PBX3, FLT3, and c-KIT). PU.1 directly binds to the genomic loci of its target genes in vivo, and is required to maintain active expression of those genes in both normal hematopoietic stem and progenitor cells and in MLL leukemia. Finally, the clinical significance of the identified PU.1 signature was indicated by its ability to predict survival in acute myelogenous leukemia patients. Together, our findings demonstrate that PU.1 contributes to the development of MLL leukemia, partially via crosstalk with the MEIS/HOX pathway.

Initiation of MLL-rearranged AML is dependent on C/EBPα

MLL-fusion proteins are potent inducers of oncogenic transformation, and their expression is considered to be the main oncogenic driving force in ∼10% of human acute myeloid leukemia (AML) patients. These oncogenic fusion proteins are responsible for the initiation of a downstream transcriptional program leading to the expression of factors such as MEIS1 and HOXA9, which in turn can replace MLL-fusion proteins in overexpression experiments. To what extent MLL fusion proteins act on their own during tumor initiation, or if they collaborate with other transcriptional regulators, is unclear. Here, we have compared gene expression profiles from human MLL-rearranged AML to normal progenitors and identified the myeloid tumor suppressor C/EBPα as a putative collaborator in MLL-rearranged AML. Interestingly, we find that deletion of Cebpa rendered murine hematopoietic progenitors completely resistant to MLL-ENL-induced leukemic transformation, whereas C/EBPα was dispensable in already established AMLs. Furthermore, we show that Cebpa-deficient granulocytic-monocytic progenitors were equally resistant to transformation and that C/EBPα collaborates with MLL-ENL in the induction of a transcriptional program, which is also apparent in human AML. Thus, our studies demonstrate a key role of C/EBPα in MLL fusion-driven transformation and find that it sharply demarcates tumor initiation and maintenance.

Epigenetic dysregulation of leukaemic HOX code in MLL‐rearranged leukaemia mouse model

HOX genes are frequently dysregulated in human leukaemia with the gene rearrangement between mixed lineage leukaemia (MLL) and partner genes. The resultant MLL fusion proteins are known to mediate leukaemia through disruption of the normal epigenetic regulation at the target gene loci. To elucidate the pathogenic role of MLL fusion proteins in HOX dysregulation in leukaemia, we generated a novel haematopoietic lineage-specific Mll-Een knock-in mouse model using a Cre-mediated inversion strategy. The Mll(Een) (/+) invertor mice developed acute myeloid leukaemia, with organomegaly of the spleen, liver and mesenteric lymph nodes caused by infiltration of blast cells. Using Mll-Een-expressing leukaemic cell lines derived from bone marrow of Mll(Een) (/+) mutant mice, we showed that induction of Hox genes in leukaemic cells was associated with hypomethylated promoter regions and an aberrant active chromatin state at the Hox loci. Knock-down of Prmt1 was insufficient to reverse the active chromatin status and the hypomethylated Hox loci, suggesting that Prmt1-mediated histone arginine methylation was only partially involved in the maintenance of Hox expression in leukaemic cells. Furthermore, in vivo analysis of bone marrow cells of Mll(Een) (/+) mice revealed a Hox expression profile similar to that of wild-type haematopoietic stem cells. The leukaemic Hox profile was highly correlated with aberrant hypomethylation of Hox promoters in the mutant mice, which highlights the importance of DNA methylation in leukaemogenic mechanisms induced by MLL fusion proteins. Our results point to the involvement of dynamic epigenetic regulations in the maintenance of the stem cell-like HOX code that initiates leukaemic stem cells in MLL-rearranged leukaemia. This provides insights for the development of alternative strategies for leukaemia treatment.

Increased Expression Of Fgf2 Induces Cytokine Hypersensitivity In Leukemia Stem Cells With Translocations Involving The MLL Gene

Leukemias with chromosomal translocation or partial tandem duplications involving the MLL (mixed lineage leukemia) gene have exceptionally poor prognosis (referred to as 11q23-leukemias). At the molecular level, 11q23-leukemias are characterized by increased expression of a set of homeodomain transcription factors in CD34+ hematopoietic stem cells (HSC) and differentiating CD34- myeloid progenitor cells. This transcription factor set includes HoxB3, B4, A7-11 and Meis1. Although many elegant studies have identified the precise molecular mechanisms by which MLL-fusion proteins increase HOX gene transcription, less is known about the downstream effects of increased Hox expression and potential implications for leukemogenesis. In prior studies, we determined that the gene encoding fibroblast growth factor 2 (Fgf2 or basic Fgf) is a HoxA10 target gene, and production of Fgf2 is increased in HoxA10 overexpressing primary bone marrow cells. In the current studies, we determine that FGF2 is also activated by HoxA9; identifying the only common target gene for HoxA9 and HoxA10 where both proteins activate transcription. We also find Mll-Ell (an MLL-fusion protein) increases FGF2 transcription and Fgf2 expression in a HoxA9 and HoxA10 dependent manner. Importantly, autocrine production of Fgf2 by Mll-Ell expressing HSC and myeloid progenitor cells increases cytokine-induced proliferation and enhances cell survival. In additional studies, we determine that these effects of Fgf2 are mediated by phospho inositol 3 kinase (PI3K) dependent stabilization of beta catenin. We find that autocrine production of Fgf2 results in increased expression of key beta catenin target genes in Mll-Ell expressing cells, including c-myc, survivin and cyclinD1. Beta catenin target genes also include the genes encoding CDX4 and HOXA10. Additionally, Cdx4 activates transcription of the HOXA9 and HOXA10 genes, and HoxA10 activates the CDX4 gene. Therefore, increased Fgf2 also establishes a positive feedback loop that enhances transcriptional activation of HOXA9 and HOXA10 by MLL-fusion proteins. 11q23-AML is generally refractory, but the molecular mechanisms for this are unknown. Our studies suggest that increased Fgf2 expression in 11q23-AML contributes to cytokine hypersensitivity and sustained survival of leukemia stem cells. Targeting the effects of Fgf2 might increase chemotherapy sensitivity in this form of leukemia. Small molecule inhibitors of Fgf-receptors are available, but have not yet been tested in this form of AML. We suggest that this would be a rational therapeutic approach to this disease. Disclosures:No relevant conflicts of interest to declare.

Regulation Of Normal and Malignant Hoxa Gene Expression Through Higher H3K79 Methylated States

The clustered Hoxa genes Hoxa5-Hoxa10 are important for self-renewal and differentiation of hematopoietic stem and progenitor cells. These genes are often aberrant upregulated in human malignancies, one of the most striking examples of which are leukemias with rearrangements of the mixed lineage leukemia (MLL) gene. Since these key posterior Hoxa cluster genes are known to show graded down-regulation upon hematopoietic differentiation, we sought to assess the epigenetic changes associated with this tightly controlled, developmentally critical transcriptional program. We performed ChIP sequencing on lineage negative, Sca-1 + Kit + (LSK) cells where the posterior Hoxa5-10genes are highly expressed compared to granulocyte macrophage progenitors (GMPs) that show markedly lower expression of these transcripts. We observed a dramatic diminution in H3K79 di/tri methylation as cells differentiated from LSKs to GMPs. In contrast, we found minimal changes in H3K79 mono-methylation during the LSK-GMP transition suggesting that high expression of Hoxa genes is maintained through higher H3K79 methylated states in normal hematopoiesis. Strikingly, the removal of H3K79 me2/3 but not me1 was sufficient for repressive epigenetic mechanisms such as polycomb repressive complex 2 (PRC2)- mediated H3K27 tri-methylation to invade the Hoxa locus during the LSK to GMP differentiation. The decrease in H3K79 di/tri methylation upon hematopoietic differentiation may be the result of a graded decrease in expression of the DOT1L co-factor Af10 which we have recently found to regulate the transition from H3K79 mono to dimethylation.Next we wanted to probe whether MLL-leukemias drive Hoxa gene expression through increased conversion of H3K79 methylation from the mono to the di/tri-methylated states. We performed ChIP sequencing for H3K79 me1/2/3 to identify the different states of H3K79 methylation across the epigenetic landscape of MLL-transformed cells. Strikingly we observed that in both murine and human leukemia, there was a dramatic hyper-conversion of H3K79 mono-methylation to di/trimethylation specifically at MLL-AF9 target genes. Af10 deletion in the MLL-AF9 leukemia significantly reduced H3K79me2/3 while retaining - and in some cases even increasing - H3K79 me1 at MLL-target genes. We then assessed chromatin accessibility at the Hoxa locus following nuclease digestion and observed that the diminution of H3K79me2/3 was enough to significantly reduce chromatin accessibility at the Hoxa locus despite the relative retention of H3K79me1. Similar to the LSK-GMP transition, reduction in H3K79 me2/3 but not H3K79me1 was accompanied by a dramatic increase in H3K27me3 in a very specific subset of genes including Hoxa5-10.The genetic and epigenetic changes resulting from Af10 deletion significantly impaired MLL-leukemogenesis and sensitized leukemia cells to DOT1L inhibition. These results suggest that the expression of Hoxa genes in hematopoiesis is controlled through higher H3K79 methylated states, possibly aided by the DOT1L co-factor Af10. The MLL-AF9 fusion seems to hijack this mode of regulation through aberrant hyperconversion of H3K79 methylation to higher methylated states at critical oncogenic targets including the Hoxa genes. These data suggest that targeting the interaction between DOT1L and Af10 is a potential therapeutic strategy since it blunts the conversion of H3K79me1 to higher states of methylation by MLL-fusion proteins and thus significantly impairs MLL-mediated transformation. Disclosures:Armstrong:Epizyme: Consultancy.

Role of Epigenetic Readers in Pathogenesis and Therapy of Acute Leukemias

Acute myeloid leukemia (AML) is an aggressive hematological malignancy with a dismal survival rate of less than 30 percent and is thus a disease where novel therapeutics are urgently required. AML is highly heterogeneous at both the molecular and clinical level. However, recurrent themes of altered transcriptional and epigenetic control are evident from sequencing and mechanistic studies. Recently, critical epigenetic regulators have surfaced as mediators of leukemogenesis and potential therapeutic targets in AML and other hematological malignancies. We and others have recently demonstrated that the Bromodomain and Extra Terminal (BET) protein family of transcriptional adapter proteins are required for transformation by MLL-fusion proteins, where they facilitate the localization of aberrant transcriptional complexes to chromatin at genes critical for leukemogenesis. Furthermore, we and others have demonstrated that small molecule inhibitors, such as I-BET151 and JQ1, of the protein-protein interactions between these proteins and acetylated lysine residues in histone tails are effective in pre-clinical models of MLL leukemias. We have recently extended studies of BET inhibition to other forms of AML using I-BET151 and found a range of sensitivities across a panel of AML cell lines and patient samples with common recurrent mutations. These cell lines also demonstrate a common cellular response to MLL-fusion leukemias, with induction of apoptosis and G1/0 arrest, suggesting that similar transcriptional programs may be targeted. In further investigating this, we identified a common transcriptional program downregulated in sensitive leukemia subtypes. In addition, using diagnostic gene expression profiles, we also demonstrate that this program can classify AML patients, suggesting that the program may provide predictive biomarkers and potentially identify those patients most likely to respond to BET inhibitions. Focusing on the most common mutation in AML, mutations of the nucleophosmin gene (NPM1c mutations), we present evidence to suggest that wild-type NPM1 has an inhibitory influence on BRD4, which is relieved upon NPM1c mutation and cytosplasmic dislocation. NPM1c mutation allows upregulation of the core transcriptional program facilitating leukemia development, and this program is abrogated by I-BET therapy. Finally, we demonstrate the efficacy of I-BET151 in human cell lines, a unique murine model, and in primary patient samples of NPM1c AML. Taken together, we have used BET inhibitors as a laboratory tool to dissect out mechanisms of aberrant transcriptional control in AML. Furthermore our pre-clinical evaluation supports clinical trials of BET inhibitors across multiple AML subtypes and identifies potential predictive biomarkers of sensitivity and response to inform these studies. Disclosures:No relevant conflicts of interest to declare.

MLL-AF9 Positive Acute Myeloid Leukemia Cells Are Sensitive To All-Trans Retinoic Acid

BackgroundAmong the subtypes of acute myeloid leukemia (AML), acute promyelocytic leukemia (APL) responds dramatically to differentiation therapy with all-trans retinoic acid (ATRA). But, ATRA is not sufficient to induce differentiation in non-APL AML. Herein, we first evaluated whether MLL fusion partners, such as MLL-AF9 and MLL-AF4/AF5q31, affect the sensitivity of ATRA in human and murine MLL fusion positive AML cell lines. In addition, we also assessed the level of H3K4me2 modification for the RARα gene in human AML cell lines, and whether the LSD1 inhibitor affected the ATRA-resistant MLLfusion positive AML cell lines. MethodsThree human AML cell lines with MLL fusion (THP1 and MOLM13 expressing MLL-AF9, and KOCL48 expressing MLL-AF4) and two murine leukemic cell lines derived from murine Lin- hematopoietic progenitors transduced by retroviral vector expressing MLL fusion genes, such as MLL-AF9 and MLL-AF5q31 were used in this study. To test the sensitivity of ATRA, all cell lines were treated with 1 μM ATRA for three days. Monocytic differentiation was assessed by morphological analysis, NBT reduction test and flow cytometric analysis (FCM) of CD11b expression. Quantitative RT-PCR (qRT-PCR) analysis and western blotting was carried out to measure the RARα, C/EBPα, C/EBPε, and PU.1 expressions. Cytotoxicity assay was performed to determine the IC50 of ATRA in these cell lines and whether ATRA could decrease the IC50 of cytarabine in MLL-AF9positive cells by using WST assays. Chromatin immunoprecipitation (ChIP) assay was performed to determine the value of H3K4me2 status using RARα-specific primers. To determine whether tranylcypromine (TCP), which is a nonreversible LSD1 inhibitor, could reactivate ATRA sensitivity, we treated KOCL48 with 10 μM TCP and 1μM ATRA. ResultsWe first determined that morphological changes characteristic of monocytic differentiation, CD11b expression and NBT reduction are more readily induced by ATRA in human and murine MLL-AF9 positive cells than MLL-AF4/AF5q31 positive cells. The NBT reduction percentage was 12.5±3.77 in THP1, 13.1±2.02 in MOLM13, but 7.00±2.64 in KOCL48 cells (p<0.05). The ATRA treatment also induced growth inhibition and increased gene expression related to monocytic differentiation through retinoic acid (RA) pathway, more efficiency in MLL-AF9 positive cells than MLL-AF4/AF5q31 cells. The IC50 of ATRA for THP1, MOLM13 and murine MLL-AF9 cells was 3.85, 1.24 and 1.95 μM, but over 10 μM for KOCL48 and murine MLL-AF5q31 cells. Furthermore, qRT-PCR and western blot revealed that ATRA increased expression level of RARα, C/EBPα, C/EBPε, and PU.1 in MLL-AF9 positive cells, but not in MLL-AF4/AF5q31 positive cells. Collectively, RA pathway is more impaired in MLL-AF4/AF5q31 positive cells than MLL-AF9 positive cells. In addition, the increase in RARα, C/EBPα, C/EBPε, and PU.1 mRNA expressions were observed in two primary MLL-AF9 positive AML cells treated with ATRA. Next, we also carried out ChIP assay and the H3K4me2/ H3 on the RARα promoter in MLL-AF9 positive cells were higher than MLL-AF4 positive cell. Furthermore, ATRA and TCP combination treatment in KOCL48 induced morphological changes, CD11b expression, and increased the expression level of RARα, C/EBPα, C/EBPε, and PU.1, suggesting that inhibition of LSD1 restores ATRA sensitivity. Finally, ATRA in combination with cytarabine treatment in MLL-AF9 positive cells enhanced cytarabine sensitivity: the IC50 of cytarabine in THP1, MOLM13, and murine MLL-AF9cells was 4.18, 0.04, and0.065 μM without ATRA and 0.13, 0.0005, and 0.015 μM with ATRA, respectively. ConclusionsOur data demonstrated that RA pathway was more profoundly impaired in MLL-AF4/AF5q31 positive cells than MLL-AF9 positive cells, suggesting type of MLL fusion protein contributes inactivation of RA pathway. Our data also identified the sensitivity of ATRA was correlated with the ratio of H3K4me2/ H3 on the RARα promoter, and TCP restore the sensitivity of ATRA in KOCL48, suggesting the decrease of the H3K4me2/H3 plays a role in inactivation of RA pathway. Additionally, we revealed that synergistic antileukemic activity of ATRA in combination with cytarabine in MLL-AF9 positive cells. Therefore, ATRA in combination with cytarabine might be novel therapeutic option for the ATRA sensitive AML cells, especially for MLL-AF9 positive cells. Disclosures:No relevant conflicts of interest to declare.

Requirement For CDK6 In MLL-Rearranged Acute Myeloid Leukemia

Chromosomal rearrangements involving the H3K4 methyltransferase MLL trigger aberrant gene expression programs in hematopoietic stem and progenitor cells and give rise to an aggressive subtype of acute myeloid leukemia (AML) that is associated with intermediate or poor survival. Insights into MLL fusion-mediated leukemogenesis have not yet translated into better therapies in the clinic, in part because mutant MLL is difficult to target directly and it is incompletely understood which of the genes affected by altered epigenetic regulation in MLL-rearranged AML are responsible for malignant transformation.To search for essential signaling pathways in MLL-rearranged AML that might serve as new therapeutic targets, we performed loss-of-function RNA interference (RNAi) screens in 5 AML cell lines (NOMO-1, THP-1, OCI-AML3, HL-60, U937) using a lentiviral short hairpin RNA (shRNA) library, and observed that the cell cycle regulator CDK6, but not its functional homolog CDK4, was preferentially required by MLL-AF9pos NOMO-1 and THP-1 cells. The enhanced CDK6 dependence of MLL-rearranged cells was confirmed in an expanded panel of AML cell lines (MLL-rearranged, n=6; MLL wildtype [WT], n=4) that also included cell lines harboring other MLL fusions (MLL-AF4 and MLL-AF6), and the RNAi-induced phenotype was countered by overexpression of an shRNA-resistant CDK6 cDNA.Stable knockdown of MLL-AF9 in MLL-AF9pos cell lines and overexpression of MLL-AF9 in WT MLL-expressing cell lines, normal human CD34pos cells, or Ba/F3 murine pro-B cells led to concordant changes in CDK6 mRNA and protein levels that resembled those of HOXA9, a known MLL-AF9 target, indicating that CDK6 is rendered essential via transcriptional activation by truncated MLL. Analysis of cell cycle distribution, apoptosis induction, and myeloid differentiation demonstrated that the differential growth-inhibitory effect of CDK6 suppression was mainly attributable to myeloid differentiation, as MLL-AF9pos cell lines upregulated CD11b expression and assumed a more mature, macrophage-like morphology upon CDK6 knockdown, effects not observed in WT MLL-expressing cells. Furthermore, the immature phenotype of NOMO-1 cells could be rescued by overexpression of an shRNA-resistant CDK6 cDNA.Consistent with the observations in AML cell lines, knockdown of Cdk6 also impaired the proliferation and in vitro clonogenic activity of primary murine bone marrow (BM) cells stably transduced with MLL-AF9, whereas cells expressing another leukemogenic fusion gene (MOZ-TIF2) and Ba/F3 cells were largely unaffected. We also expressed MLL-AF9 in unfractionated BM derived from Cdk6 knockout mice and observed that colony numbers were gradually reduced in cultures initiated with Cdk6+/- and Cdk6-/- BM compared to WT BM. Furthermore, most of the colonies obtained were small and displayed loose morphology in contrast to the large, dense, blast-like colonies seen in cultures initiated with transduced WT BM. We are currently investigating whether Cdk6 is also required for AML development and propagation in vivo using a murine BM transplantation model of MLL-AF9-induced leukemia.The context-dependent effects of lowering CDK6 expression could be recapitulated in cell lines and primary human AML specimens using palbociclib (also known as PD-0332991), a small-molecule inhibitor of CDK4 and CDK6 enzymatic activity that is in clinical development as an anticancer agent. We are currently devising strategies to combine this compound with cytotoxic chemotherapy as well as other targeted therapeutics, such as small-molecule bromodomain inhibitors, to maximize killing of MLL-rearranged AML cells.Together, our data (1) identify CDK6 as a critical and potentially “actionable” effector of MLL fusion proteins in leukemogenesis, (2) link the catalytic activity of CDK6 to arrested myeloid differentiation in MLL-rearranged AML, and (3) underscore that cell cycle regulators thought to normally act redundantly may have distinct functions in different genetic contexts. Disclosures:No relevant conflicts of interest to declare.