Search

Lymphoma with secondary central nervous system (CNS) involvement represents one of the most aggressive malignancies, with poor prognosis and high mortality. New diagnostic tools for its early detection, response evaluation, and CNS relapse prediction are needed. We analyzed circulating microRNAs in the cerebrospinal fluid (CSF) and plasma of 162 patients with aggressive B-cell non-Hodgkin’s lymphomas (B-NHL) and compared their levels in CNS-involving lymphomas versus in systemic lymphomas, at diagnosis and during treatment and CNS relapse. We identified a set of five oncogenic microRNAs (miR-19a, miR-20a, miR-21, miR-92a, and miR-155) in CSF that detect, with high sensitivity, secondary CNS lymphoma involvement in aggressive B-NHL, including DLBCL, MCL, and Burkitt lymphoma. Their combination into an oncomiR index enables the separation of CNS lymphomas from systemic lymphomas or nonmalignant controls with high sensitivity and specificity, and high Receiver Operating Characteristics (DLBCL AUC = 0.96, MCL = 0.93, BL = 1.0). Longitudinal analysis showed that oncomiR levels reflect treatment efficacy and clinical outcomes, allowing their monitoring and prediction. In contrast to conventional methods, CSF oncomiRs enable detection of early and residual CNS involvement, as well as parenchymal involvement. These circulating oncomiRs increase 1–4 months before CNS relapse, allowing its early detection and improving the prediction of CNS relapse risk in DLBCL. Similar effects were detectable, to a lesser extent, in plasma.

Precise levels of transcription factor PU.1 are critical for differentiation to distinct blood lineages, and even modest decreases leads to leukemogenesis. However, the mechanism of how target genes sense different PU.1 levels is unknown. To address this, we used myeloid progenitors encoding inducible PU.1 transgene. Intermediate PU.1 activity induced differentiation into granulocyte-like cells, while high PU.1 produced macrophages. Comparative analysis of regulatory sequences (+-50kb) of two genes activated exclusively by intermediate PU.1, Myeloperoxidase (Mpo) and Gelatinase B (Mmp9) identified 14 Mpo and 16 Mmp9 putative enhancers resembling superenhancers. Interestingly, the PU.1 binding motifs within these regions have an intermediate to high affinity (log ods score). To test their functionality we assembled successively these regulatory sequences into luciferase vectors. Specifically, we identified novel enhancer elements in -3.4 and -15 kb of Mpo which responded to intermediate (but not high) PU.1 levels. Interestingly, activity of -3.4 kb enhancer required presence of promoter while the -15kb element required presence of both PP and the -3.4kb element. Similar phenomenon was observed at multiple Mmp9 enhancers. This model is supported by our Chromosome confirmation capture (3C) data. Native activity of tested enhancers was confirmed by expression of bidirectional enhancer RNAs. The mutation of affinity of PU.1 binding sites within enhancers lead to reprogramming of their activity. In conclusion, our data support the model that PU.1 at intermediate concentrations binds to high affinity binding sites in several enhancers of granulocyte genes, causing their successive interaction that leads to transcription activation. Collectively, this suggests that a cooperative assembly of several cell-type specific enhancers is required for optimal Mpo and Mmp9 activation.

The oncogenic cluster miR-17-92 encodes seven related microRNAs that regulate cell proliferation, apoptosis and development. Expression of miR-17-92 cluster is decreased upon cell differentiation. Here, we report a novel mechanism of the regulation of miR-17-92 cluster. Using transgenic PU.1−/− myeloid progenitors we show that upon macrophage differentiation, the transcription factor PU.1 induces the secondary determinant Egr2 which, in turn, directly represses miR-17-92 expression by recruiting histone demethylase Jarid1b leading to histone H3 lysine K4 demethylation within the CpG island at the miR-17-92 promoter. Conversely, Egr2 itself is targeted by miR-17-92, indicating existence of mutual regulatory relationship between miR-17-92 and Egr2. Furthermore, restoring EGR2 levels in primary acute myeloid leukaemia blasts expressing elevated levels of miR-17-92 and low levels of PU.1 and EGR2 leads to downregulation of miR-17-92 and restored expression of its targets p21CIP1 and BIM. We propose that upon macrophage differentiation PU.1 represses the miR-17-92 cluster promoter by an Egr-2/Jarid1b-mediated H3K4 demethylation mechanism whose deregulation may contribute to leukaemic states.

Elevated levels of microRNA miR-155 represent a candidate pathogenic factor in chronic B-lymphocytic leukemia (B-CLL). In this study, we present evidence that MYB (v-myb myeloblastosis viral oncogene homolog) is overexpressed in a subset of B-CLL patients. MYB physically associates with the promoter of miR-155 host gene (MIR155HG, also known as BIC, B-cell integration cluster) and stimulates its transcription. This coincides with the hypermethylated histone H3K4 residue and spread hyperacetylation of H3K9 at MIR155HG promoter. Our data provide evidence of oncogenic activities of MYB in B-CLL that include its stimulatory role in MIR155HG transcription.

Introduction: Currently, there is no consensus regarding diagnosis and treatment of occult lymphomatous menigeal involvement of cerebrospinal fluid (CSF) in patients with systemic non-Hodgkin lymphomas (NHL), especially in patients with minimal number of clonal cells detected by flow cytometry (FCM) or by cytology. Aim of the study: To describe a cohort of patients with occult lymphomatous involvement in cerebrospinal fluid including the diagnostic methods as well as management. Patients and methods: CSF at diagnosis of B-NHL was examined by FCM, cytology, biochemistry and microRNA analysis in a cohort of 62 patients (systemic DLBCL N=45, MCL N=15, Burkitt lymphoma N=2) between 2010 and 2013. Occult meningeal involvement was defined by: a/ FCM: absolute positive number of clonal cells Results and conclusions: According to described criteria we identified 8 patients with occult meningeal involvement at diagnosis (systemic DLBCL N=4, MCL N=3, Burkitt lymphoma N=1) out of 62 patients: 3 with negative FCM and positive cytology (2-14 cells/ul); and 5 with positive FCM These results suggest that examination of oncogenic microRNA in CSF at diagnosis is more sensitive method than FCM and cytology of CSF to detect the patients with occult CNS involvement and higher risk of CNS involvement and this should be reflected in treatment strategy in further clinical trials. Grants: GACR305 13-12449P, UNCE204021, PRVOUK: P24/LF1/3, P27/LF1/1, P 27/2012 LF3 Disclosures Spacek: Roche: Consultancy, Travel grants Other.

Introduction MicroRNAs (miRNAs) are single-stranded non-coding RNAs (21-25nt) that negatively regulate gene expression at post-transcriptional level by blocking translation and mRNA destabilization. Diffuse large B cell lymphomas (DLBCL) overexpress oncogenic microRNAs (including miR-17-92 and miR-155) that regulate tumor growth and spreading. Unlike longer species of RNA, microRNAs are stable and detectable in body fluids including serum/plasma and cerebrospinal fluid (CSF), however their kinetics in body fluids upon course the therapy of DLBCL has not been fully explored. Notably DLBCLs with primary (PCNSL) or secondary CNS involvement display poor prognosis, short overall survival and their current diagnosis and disease monitoring is based on conventiontional methods (magnetic resonance imaging, histopathology, CSF cytology and flow cytometry) with unsatisfactory sensitivity. Aim We investigated abundance of microRNAs in the CSF and sera of DLBCL patients with (N=18) or without (N=30) CNS involvement and compared their levels at diagnosis and during treatment-response phase. Methods Extensive time-points to collect CSF and sera (5-20) were used during and after administration of therapy. Twenty candidate miRs from DLBCL patients were quantitated by TaqMan RT-PCR and their levels adjusted to those of control miRs and normal donors. Results Analysis of CSF at diagnosis revealed that most prominently deregulated miRs (members of oncogenic miR-17-92 cluster and miR-21) were specificaly elevated in both PCNSL (N=11) and systemic lymphomas with CNS involvement (SCNSL, N=7) to similar extent while in systemic lymphomas without CNS involvement (N=30) their levels were significantly lower. Analysis of sera at diagnosis revealed that all SCNSL patients and ∼50% of systemic lymphomas without CNS involvement expressed high levels of the oncomiRs. Surprisingly, also a subset of PCNSL patients had elevated these miRs in the sera and in turn, a subset of systemic lymphomas without CNS involvement showed moderately elevated oncomiRs in the CSF. These data indicate that 1) lymphoma patients with CNS involvement (PCNSL or SCNSL) are characterized by increase of oncogenic microRNAs in CSF, 2) blood brain barrier is likely permissive for miRs in both directions. Next, we investigated the dynamics of abundance of the oncomiRs in CSF upon course of the treatment in 8 patients with PCNSL or SCNSL. In therapy-responders (N=4) the oncomiRs gradually decreased in CSF. However, their levels were not completely normalized as compared to normal individuals. In contrast, refractory lymphomas (N=2) displayed gradual increase of these miRs in CSF regardless the treatment. Interestingly, clinical relapse within CNS as confirmed by MRI and cytology (N=2) was preceded (>3 months) by significant and gradual elevation of miRs in CSF. In systemic lymphoma (N=2) increased levels of studied oncomiRs in CSF preceded the CNS progression. In all cases high levels of miRs in CSF preceded and positively correlated with data from flow cytometry and cytology. Conclusions The measurement of oncomiRs in CSF and sera represents sensitive tool for detection of CNS lymphoma, for monitoring and estimation of therapy efficacy and prediction of the CNS relapse. Furthermore, our data indicate that elevated expression of oncomiRs in CSF and sera in systemic lymphomas may become helpful to predict CNS lymphoma involvement. Grants: GACR305 13-12449P, UNCE204021, PRVOUK: P24/LF1/3, P27/LF1/1, P 27/2012 LF3 Disclosures: Trneny: Roche: Honoraria, Research Funding.

Abstract 4624 Mantle cell lymphoma (MCL) represents B-cell lymphoma derived from the mantle zone that surrounds normal germinal center follicles. Pathophysiology of this hardly curable disease involves t(11,14)(q13,q32) translocation which leads to upregulation of Cyclin D1(CCND1). Recently, microRNAs were demonstrated to significantly modify MCL pathogenesis (Jian-Jun Zhao et al. 2010) and therapy responsiveness (Jiang et al. 2010). In order to broaden our knowledge of regulatory pathways in MCL we searched for differentially expressed microRNAs and their differentially expressed mRNA targets between MCL samples and control samples. Samples consist of magnetically separated B cells derived from peripheral blood. We used 1) Microarray mRNA hybridization [Affymetrix Human Genome U133 Plus 2.0 Array, N(MCL)=5, N(control)=5] and 2) microRNA profiling [TaqMan® Array Human MicroRNA Card A v2.0 technology, N(MCL)=5, N(control)=5] followed by statistical analysis [limma (Smyth 2005)]. Differentially regulated targets of the deregulated microRNAs were selected from the databases involving both predicted (Betel et al. 2007) or confirmed (Hsu et al. 2010) target mRNAs. Among the most significant upregulated microRNAs (exceeding 10 fold) are miR-9, miR-124 and miR-183. We have found that upregulated miR-9 has confirmed downregulated target gene PRDM1 (PR domain zinc finger protein 1) which plays a role in B cell maturation (Turner et al. 1994) and may act as a tumor suppressor (Pasqualucci et al. 2006). Upregulation of miR-9 and downregulation of its targets was recently demonstrated in Burkitt lymphoma (Onnis A et al. 2010) and Hodgkin lymphoma (Nie K et al. 2008). Two additional upregulated microRNAs, miR-124 and miR-183, yet not associated with lymphomas, have downregulated target gene Integrin beta-1 (CD29) which regulates survival (Fukumori et al. 2003). Among most significant downregulated microRNAs is miR-101 (FC=-7). Downregulated microRNA miR-101 has known oncogene N-Myc (MYCN) as upregulated confirmed target and DNA (cytosine-5)-methyltransferase 3A (DNMT3A) as upregulated target. Overexpression of a well known hematopoietic oncogene MYCN and epigenetic repressor DNMT3A may represent additional pathogenesis-related factors in MCL. Our data support importance of the candidate mechanisms involving microRNAs and their target programs in pathogenesis of MCL. They are currently extended on a larger patient cohort by analyzing expression and functional significance of the candidate microRNAs. (Grants: NT10310-3/2009, MPO FR-TI2/509, NPVII 2B06077, MSM 0021620806, LC 06044, SVV-2011-262507). Disclosures: No relevant conflicts of interest to declare.

Abstract 3663 MicroRNAs are non-coding RNAs (19–23nt) that negatively regulate expression by interference with the translation or RNA stability of up to hundreds of targets. Tumors including leukemias and lymphomas are associated with deregulated expression of microRNAs (Volinia et al. 2010). Several microRNAs including miR-17-92 cluster members and miR-155 are overexpressed in lymphomas and showed promise in predicting relapse (Fabbri M et al. 2011). Unlike longer species of RNA, microRNAs are preserved and detectable in human fluids including serum, plasma and cerebrospinal fluid (CSF) (Baraniskin A. et al. 2011). We investigated abundance of microRNAs in the CSF and sera and tested whether their levels are associated with diagnosis or treatment of lymphomas. We studied patients with primary central nervous system lymphomas (PCNSL) (N=3) and systemic lymphomas either with CNS dissemination (N=3) or without (N=11). Histological type was predominantly diffuse large B cell lymphoma - DLBCL (N=13). Sera of a control group were also used. 200 uL of CSF and sera (both cell free) were used for total RNA isolation followed by PCR and the data were adjusted to levels of control microRNAs (miR-let-7a or miR-24). Our data demonstrate that the microRNAs of the miR-17-92 cluster (miR-19a, miR-20a, miR-92a) and miR-106b-25 cluster (miR-106b, miR-25) and miR-155 are detectable (unlike miR-106a, miR-18a) in both the CSF and sera. Particularly, miR-17-92 and miR-106b-25 are increased in CSF of the PCNSL and the systemic lymphomas with CNS involvement compared with systemic lymphomas without CNS involvement. Conversely, in the sera, expression of miR-17-92 is increased in systemic lymphomas compared to PCNSL and the control group. Next, we investigated the levels of these microRNAs in a patient with systemic DLBCL with CNS dissemination during the treatment (R-CHOP alternating with RMPV) resulting in complete remission. Analysis of the CSF and also of the sera (at 9 different time points within three months) revealed a gradual decrease (∼6 fold) of the levels of miR-19a, miR-20a, miR-92a, miR-106b, and miR-25. However, the levels of these microRNAs in the sera at the stage of complete remission were still 2 fold higher compared to controls. We conclude that measurement of microRNAs in both CSF and sera represents novel sensitive tool, additional to other approaches, for tumor detection and/or for estimation of the therapy efficacy. (Grants: NT10310-3/2009, MPO FR-TI2/509, NPVII 2B06077, MSM 0021620806, LC 06044, SVV-2011-262507) Disclosures: No relevant conflicts of interest to declare.

Abstract 4600 MicroRNA miR-155 represents a candidate pathogenic factor in chronic lymphocytic leukemia (CLL) and lymphomas (Eis PS, et al. 2005, Fulci V, et al. 2007, Calin GA, et al. 2005, Kluiver J, et al. 2005, Metzler M, et al. 2004, van den Berg A, et al. 2003., Vargova K. et al 2011). In diffuse large B-cell lymphoma (DLBCL) expression of miR-155 was associated with NFkB pathway (Rai et al., 2008). In addition, increased miR-155 levels were associated with more aggressive (Activated B-cell like) DLBCL (Eis et al. 2005). The targets of up-regulated miR-155 apparently include a key hematopoietic transcription factor PU.1. Down-regulation of PU.1 represents an important and critical step for both myeloid as well as lymphoid tumorigenesis. We herein found the increased levels of miR-155 coupled with downregulation of its target PU.1 are frequent events in B cells of a vast majority of CLL patients (N=169). To advance understanding of miR-155 and PU.1 in lymphoid malignancies we studied lymph node biopsies derived from patients with lymphomas. We show that increased miR-155 levels are also found in DLBCL (N=31, P < 0.0001), Hodgkin lymphoma (HL, N=22, P < 0.0001), follicular lymphoma (FL, N=23, P < 0.001), small lymphocytic lymphoma (SLL, N=12, P < 0.01), and marginal zone lymphoma (MZL, N=11, P < 0.01). The miR-155-overexpressing tumors display downregulation of PU.1. In contrast, peripheral T cell lymphoma (T-NHL, N=6) and mantle cell lymphoma (MCL, N=7) expressed miR-155 and PU.1 comparably as control lymph nodes. Our data indicate that expression pattern of miR-155 and its target PU.1 represents a hallmark of some but not all lymphoid malignancies. Interestingly, the miR-155/PU1 levels vary substantially within each diagnosis, therefore, our aim is to further analyze the occurrence of this relationship in different prognostic groups and analyze the clinical outcome as well. (Grants: NT10310-3/2009 & NT11218-6/2010, MPO FR-TI2/509, GAUK251135 82210 & 251070 45410, SVV-2011-262507) Disclosures: No relevant conflicts of interest to declare.

Abstract 3589 Elevated levels of microRNA miR-155 represent a candidate pathogenic factor in chronic B-lymphocytic leukemia (B-CLL). In our study, B-CLL patients (N=71, F/M 31/40, median age 65) were subject to molecular analyses that found significant overexpression of MYB and miR-155 in B-CLL and physical association of MYB with the promoter of MIR155 host gene (MIR155HG, also known as B-cell integration cluster). In CD19+ B-cells derived from healthy control individuals (N=13), MYB and miR-155 expression was significantly lower. Next, we found that MYB positively regulates MIR155HG transcription in reporter assays. The endogenous chromatin structure of the MIR155HG promoter in B-CLL cells is characterized by spreading of active chromatin mark, histone H3K9 acetylation. Gene expression arrays on B-CLL patients (N=11) identified a set of predicted miR-155 target genes (N=94) that are downregulated. Their expression pattern displayed significant negative correlation with pri-miR-155 levels. Similarly, a number of MYB target genes were found deregulated (N=99) in B-CLL. Gene annotation of differentially expressed miR-155 and MYB targets in B-CLL revealed their biological functions as regulators of apoptosis (BCL2, API5), proliferation (CCND1, CCND2) and mediators of B-cell function (CD200, F11R). In addition, expression patterns of miR-155, MYB and their targets are currently being compared in a larger patient group with prognostic hallmarks of CLL (ZAP70, CD38, del(17)(p13.1), and IgVH). Our data collectively support novel candidate mechanism in B-CLL that includes MYB directly stimulating MIR155HG promoter coincident with its epigenetic dysregulation. (Grant # IGA 10310-3, MSMT 2B06077, 0021620806, LC06044, SVV-2010-254260507). Disclosures: Trneny: ROCHE: Honoraria, Research Funding.

Abstract 390 The miR-17-92 cluster (Oncomir1) encodes seven related microRNAs associated with cell proliferation, apoptosis and development and is overexpressed in number of malignancies including myeloid leukemias. The miR-17-92 cluster is highly expressed in myeloid progenitors, while it becomes downregulated upon the onset of macrophage differentiation. Conversely, sustained expression of miR-17-92 is associated with a differentiation blockade (Fontana 2007). Here we report a novel mechanism of the regulation of miR-17-92 cluster within differentiating myeloid progenitors. During macrophage differentiation, the myeloid transcription factor PU.1 transcriptionally induces the secondary determinant Early growth factor 2 (Egr2). Subsequently, Egr2 binds to the miR-17-92 cluster promoter and recruits histone demethylase Jarid1b resulting in histone H3 lysine K4 (H3K4) demethylation within the CpG island upstream miR-17-92 cluster leading to the repression of miR-17-92 expression. Reporter assays using deletion constructs of the miR-17-92 promoter region revealed that Egr2 is required for targeting the Jarid1b onto critical minimal region within CpG island of upstream miR-17-92 locus and its decreased H3K4 methylation. In addition, functional assays identified the 3'UTR of Egr2 as the target of multiple miRNAs of the miR-17-92 cluster, indicating existence of a mutual regulation between miR-17-92 cluster and Egr2, putatively involved in macrophage differentiation characterized by a bistable state, where Egr2 negatively regulates miR-17-92 cluster in differentiating cells and, in turn, miR-17-92 cluster negatively regulates Egr2 in highly proliferating progenitor cells to achieve homeostatic regulation. Ectopic expression of miR-17-92 cluster within myeloid progenitors blocked macrophage differentiation indicating that leukemogenesis may involve miR-17-92 mediated differentiation blockade. To determine if the newly identified negative regulatory mechanism of miR-17-92 is involved in leukemogenesis, we tested peripheral blood mononuclear cells isolated from acute myeloid leukemia (AML) patients (N=27). 14 of 27 AML patients exhibited significantly downregulated transcriptional factors PU.1 and Egr2 and high levels of the miR-17-92 cluster expression. Ectopic expression of Egr2 within AML blast cells (N=2) led to decreased levels of miR-17-92 and restored the expression of p21 and BIM, two established miR-17-92 targets downregulated in AML. We conclude that PU.1 represses miR-17-92 cluster during macrophage differentiation by Egr2 mediated recruitment of the Jarid1b demethylase and that dysregulation in the miR-17-92 repression mechanism may contribute to the pathogenesis of AML. (Grants # IGA 10310-3, MSMT 2B06077, 0021620806, LC06044, SVV-2010-254260507). Disclosures: No relevant conflicts of interest to declare.

PU.1 (Sfpi1) is an ets family transcription factor required for the proper generation of both myeloid and lymphoid lineages. Distinct nuclear levels of PU.1 are required for macrophage differentiation (DeKoter 2000). MicroRNAs are a novel class of 22nt long non-coding post-transcriptional regulators, associated with differentiation, development and carcinogenesis. To identify miRNAs involved in PU.1 dependent hematopoietic differentiation we used PU.1-null hematopoietic progenitor cells stably expressing PU.1 fused to the estrogen receptor ligand-binding domain inducible by tamoxifen (PUER) (Walsh 2002). PU.1 affected expression approximately 100 mature miRNA. Among them, we have observed that PU.1 downregulates members of miR-17~92 cluster and its paralog miR-106b~25 previously associated with macrophage differentiation, apoptosis and hematologic and other malignancies (Ota 2004, Hayashita 2005, Fontana 2008). During macrophage differentiation miR-17~92 becomes downregulated and its sustained expression is associated with the differentiation block (Fontana 2007). In tamoxifen-stimulated PUER cells, PU.1 efficiently downregulates mature miRNAs of the miR-17~92 (-17.5p, -20a, -92) cluster and its paralog miR-106b~25 (but not miR-106a~363) within 96 hours and in dose dependent manner. PU.1 in the same cells activates expression of Egr2, myeloid transcription factor promoting macrophage transcriptional program a putative 8-mer-seed sequence-containing target of miR-17~92 cluster (predicted by TargetScan). We tested whether PU.1 represses expression of miR-17~92 cluster by inducing Egr2. We demonstrate that ectopic expression of Egr2 in unstimulated PUER cells inhibited expression of miR-17~92 cluster, both at the levels of mature miRs and pri-miRNAs, and induced expression of macrophage differentiation markers (Cd14, Csf1r) coupled with cell cycle arrest. Trichostatin-dependent derepression of miR-17~92 indicated that its expression is regulated at the chromatin level. Chromatin immunoprecipitation revealed specific Egr2 occupancy within the regulatory region upstream of the miR-17~92 cluster. Our data indicate that in differentiating cells, a mechanism exists involving direct repression of the miR-17~92 cluster that in turn has the potential to inhibit Egr2 expression. We have tested this mechanisms in peripheral blood mononuclear cells isolated from acute myeloid leukemia (AML) patients (N=9) where levels of PU.1 and Egr2 are significantly lower that in normal controls. 5 out of 9 AML patients significantly overexpressed miR- 17.5p, miR-20a, miR-92 coinciding with downregulation of its established targets p21 and BIM. Rescue of Egr2 in AML cells (N=2) led to downregulation of miR-17~92 cluster and upregulation of BIM mRNA levels coincident with induction of apoptosis. We conclude that PU.1 regulates transcription of the miR-17~92 cluster by inducing Egr2 and demonstrate this mechanism in differentiating macrophages and its disruption in AML cells.

PU.1 (Sfpi1) is an ets family transcription factor required for the proper generation of both myeloid (macrophages and neutrophils) and lymphoid lineages (B and T lymphocytes)(Scott 1994, McKercher 1996). Graded expression of exogenous PU.1 in murine PU.1-deficient fetal liver hematopoietic progenitors demonstrated that increased levels of PU.1 are required to initiate development of macrophages (DeKoter, 2000). We have studied the effects of graded expression of PU.1 on its occupancy in chromatin and on the development of myeloid cells in vitro. We measured changes in gene expression, PU.1 occupancy and histone modifications in PU.1-null hematopoietic progenitor cells stably expressing PU.1 fused to the ligand-binding domain of the estrogen receptor (PU.1-ER) (Walsh 2002). The level of active PU.1-ER was regulated with graded levels of the ER inducer tamoxifen. In vitro, intermediate levels of tamoxifen produced cells with granulocyte characteristics in the suspension cell fraction and macrophage-like characteristics in the attached fraction, whereas high levels of PU.1 produced mostly attached macrophage-like cells. Expression of granulocyte-specific PU.1 target mRNAs including gelatinase B (Mmp9) and myeloperoxidase (Mpo) were observed to be expressed only with intermediate levels of tamoxifen. In contrast, expression of macrophage PU.1 target mRNAs including Cd14, F4/80 and Cd68 mRNAs were observed to be gradually upregulated upon PU.1-ER activation, with the maximum expression at the highest levels of tamoxifen. Thus, the expression levels of PU.1 target genes and phenotypic characteristics of the cells are dependent on PU.1 levels. Interestingly, macrophage-like cells can be produced from granulocytic-like cells by changing tamoxifen levels and vice versa. Chromatin immunoprecipitation analysis revealed specific PU.1 occupancy within regulatory regions of the genes predominantly expressed in macrophages including Cd14 and Cd11b after treatment with high levels of tamoxifen. Specific PU.1 occupancy within regulatory regions of the granulocyte specific genes including MMP9 was observed at intermediate levels of tamoxifen. Suprisingly, chromatin immunoprecipitation analysis revealed specific PU.1 occupancy within regulatory regions of the lymphocytic PU.1 target genes including Interleukin-7 receptor (Il-7r) and RAG1 at intermediate levels of tamoxifen even though expression of these genes was not detected. Accumulation of acetylated K9 and methylated K4 of histone H3 in gene loci of macrophage and granulocytic markers such as Cd14, Cd11b, and Mmp9 correlated with their mRNA expression. However, lymphocyte-specific regulatory regions including that of Il-7r gene were hypoacetylated in H3K9 despite a marked PU.1 recruitment suggesting additional factors may be required for PU.1 mediated transactivation. To identify these molecules we have tested PU.1-dependent transcription factors: Egr2, Nab2, Cebpa and Gfi-1 and found that upon increasing PU.1 levels, expression of Egr2/Nab2 and Gfi-1/Cebpa changed in a reciprocal manner and these changes preceded expression of the lineage specific markers. We are currently testing if PU.1 directly regulates expression of Egr2, Nab2, Cebpa and Gfi-1 during granulocytic/macrophage differentiation.

MicroRNAs are non-coding post-transcriptional regulators of gene expression often dysregulated in human cancers including hematological malignancies. MiR-155 is a putative oncogenic factor found overexpressed in chronic lymphocytic leukemia (CLL) cells. We have confirmed the increased levels of miR-155 in peripheral blood lymphocytes in 12 patients with chronic lymphocytic leukemia (CLL) compared to 11 normal controls, using quantitative PCR assay. Among the predicted high-scoring hematopoietic target genes of miR-155 is the ets-family transcription factor PU.1, a key regulator of myelo-lymphoid lineage development. Using gene expression profiling of CLL peripheral lymphocyte samples compared to normal controls, we have determined significant downregulation of the PU.1 mRNA and of additional numerous miR-155 targets previously reported to be required for myelo-lymphoid differentiation, and of a set of gene targets of PU.1-chromatin DNA interactions that we determined by chromatin immunoprecipitation analyses. Functional pathway analysis of the miR-155 targets downregulated in CLL patients suggest the suppression of T-cell receptor pathway, TGF-b/Bmp, MAPK, Jak/Stat and Wnt signaling and remodeling of actin cytoskeleton and cell-matrix interactions, as well as the regulation of numerous transcription regulators including c-FOS, c-MYC, MAFB and HDAC4. Recent reports suggested a role of miR-155 in lymphocytic differentiation, we have tested whether PU.1 may regulate expression of its own regulator miR-155. We used two cell systems expressing a transgene encoding PU.1 fused with the ligand binding domain of the estrogen receptor, ER): murine erythroleukemia (MEL) cells (Rao 1997) and PU.1 null hematopoietic progenitor cells stably transfected with PU.1-ER (Walsh 2002). Using a miRNA microarray assay we have identified a set of 21 miRNAs that are modulated by PU.1. Indeed, miR-155 was among the upregulated targets alongside a known PU.1 miR target miR-233. The direct interaction between miR-155 promoter and PU.1 protein was confirmed by chromatin immunoprecipitation in MEL cells. In summary, our data suggests that increased miR-155 levels may contribute to CLL by inhibiting PU.1 function and suppressing PU.1-mediated programs and that in two independent conditional PU.1 cell systems PU.1 directly binds miR-155 promoter and upregulates its expression suggesting thus a mutual regulatory mechanism between PU.1 and miR-155.

PU.1 and GATA1 are hematopoietic lineage-specific transcription factors that play key roles in normal myeloid and erythroid differentiation respectively. Inappropriate expression of PU.1 in proerythroblasts causes its binding to GATA-1 on DNA resulting in a block of erythroid differentiation and development of murine erythroleukemia (MEL) (Rekhtman 1999). Activation of a conditional transgene of GATA-1 (fused with the ligand binding domain of the estrogen receptor, ER) in MEL cells disrupts PU.1-mediated repression in chromatin leading to re-intiation of erythroid differentiation and cell cycle arrest (Choe 2003, Stopka 2005). In this study we show that MEL cells can also be induced to express myeloid differentiation programs upon PU.1-ER activation. Gene expression microarray analysis of GATA-1-ER MEL cells and PU.1-ER MEL cells treated with ER activators allowed us to identify mRNAs that are regulated by both GATA-1 and PU.1 including the set of GATA-1 targets repressed by PU.1, as well as the set of PU.1 targets repressed by GATA-1 in MEL cells. The targets of mutual antagonism of PU.1 and GATA-1 consisted of lineage specific transcription factors, differentiation markers and genes that cause cell cycle arrest and antiapoptotic regulators previously associated with myeloid and erythroid cell differentiation. To determine if PU.1 and GATA-1 directly regulate the lineage specific transcription factor genes, we performed chromatin immunoprecipitation (ChIP) and analyzed the ChIP samples on microarrays and by qPCR. We found that PU.1 and GATA-1 are localized near PU.1 binding sites in the genes for myeloid transcription factors Cebpa and Cbfb and near GATA-1 binding sites in the genes for erythroid transcription factors Fog1 and Nfe2. In addition, further ChIP experiments delineated chromatin architecture near the binding sites for PU.1 and GATA-1 including histone H3 content and acetylation of histone H3K9. We propose that the mutual antagonism of PU.1 and GATA-1 in inhibiting the respective differentiation programs is rendered through specific changes in chromatin structure around lineage specific transcription factors genes. These changes may contribute to the block to differentiation evident in leukemias.

Mantle cell lymphoma (MCL) is a relatively distinct B-cell non-Hodgkin lymphoma subtype with aggressive and often recurrent clinical course. At diagnosis, MCL often manifests with leukemization, a feature more common to chronic lymphocytic leukemia (CLL). Common features and differences between MCL and CLL were not yet explored by comprehensive global approaches, despite such understanding potentially being very neat for deciphering pathogenesis and tailoring therapies of these clinically distinct diseases. In our study, we have compared MCL(n=10), CLL(n=10) and normal control(n=8) B-cell samples using the Affymetrix Human Genome HG-U133 Plus 2.0 Array. We studied different mRNA levels of ~47.000 transcripts represented on the array. The comparative analyses identified a set of 892 differentially expressed genes between MCL and NBC; and 774 differentially expressed genes between CLL and NBC. In order to find MCL/CLL-specific biomarkers we focused on the intersection of differently expressed genes in both groups (CLL vs NBC and MCL vs NBC). There were 222 mRNAs in the intersection, 216 of them were deregulated in the same direction in both groups while 6 mRNAs were deregulated in the opposite direction. This set of 6 disease-specific mRNAs contained previously reported biomarkers (CD200, LEF1), and also the Myristoylated alanine-rich C-kinase substrate (MARCKS) that has not yet been studied in MCL. Thus we utilized the validation patient groups (NMCL=6, NCLL=8) and confirmed differential expression of MARCKS on protein levels by flow cytometry and immunofluorescence. As MARCKS was previously shown to either bound to the cell membrane, to reside in the cytosol, or alternatively become transmitted to nuclei, we investigated subcellular localization of MARCKS using immunofluorescence (IF). The cytoplasmic MARCKS signal in MCL was significantly higher than in CLL while the opposite was observed for the nuclear IF signal. The ratio between cytoplasmic and nuclear signal was 2.5 for MCL and 0.8 for CLL (p < 0,0001). The active forms of MARCKS were shown to become phosphorylated on serineresidues and this prompted us to study the phosphorylation forms of MARCKS in MCL. Indeed, one of the residues, Ser159/163, was hyperphosphorylated in the MCL cytoplasm and its level and distribution markedly differed from CLL or NBC. We next searched for regulatory mechanisms upstream of the MARCKS expression in MCL vs CLL. MARCKS is a predicted target of several microRNAs (according to DIANA-TarBase v7.0), among them also of miR-155 (that is differentially expressed between MCL and CLL). To further investigate the regulatory relationship between mir-155 and MARCKS we utilized a CLL cell line MEC-1 and using the CRISPR/Cas9 technology we prepared individual cell clones that were mutated within the mature miR-155 sequence that recognizes MARCKS mRNA. As expected, the miR-155-MEC-1 mutants expressed markedly higher level of MARCKS compared to the control MEC-1 cells. In conclusion, our work identified a set of six differentially expressed mRNAs when comparing MCL and CLL, among them, MARCKS. We further showed that MARCKS is differentially expressed, localized, and phosphorylated between MCL and CLL, and that MARCKS is partly controlled by oncogenic microRNA miR-155. MARCKS may play an important role in MCL pathogenesis and can serve as useful MCL biomarker. Grant support: GAČR 16-05649S & P305/12/1033, AZV: 16-27790A and 16-31586A. Institutional support: CZ.1.05/1.1.00/02.0109, UNCE 204021, LH15170, PRVOUK P24, LQ1604. Disclosures No relevant conflicts of interest to declare.

Abstract 3791 Use of DNA demethylating agent 5-azacytidine (AZA) in high risk Myelodysplastic syndromes (MDS) represents an important strategy yet not fully understood at molecular levels. Previous work demonstrated that normal progenitors treated by DNA demethylation agent Decitabine upregulated a key hematopoietic transcription factor PU.1. This was potentiated in presence of granulocyte colony stimulating factor (G-CSF) prior the DNA demethylation (Hu et al. 2010). Similarly, Decitabine administered to multiple myeloma cells resulted in DNA demethylation and upregulation of PU.1 gene (Tatetsu et al. 2007). To advance our understanding of molecular events mediated by AZA in MDS we measured effect of AZA on transformed cell lines (OCI-M2, SKM-1, MOLM-13, HNT-34, F-36P) and monitored the hematopoietic transcriptional program. We herein present evidence that AZA, while inducing efficiently cell cycle arrest and apoptosis of MDS tumor cells, stimulates also the expression of PU.1 and its target C/EBPa that is followed by acquisition of the myeloid differentiation program. Gene derepression of PU.1 is marked by epigenetic modifications at the upstream regulatory element (URE) of PU.1. G-CSF potentiated effects of AZA in MOLM-13 and SKM-1 cells. The effects of G-CSF and other cytokines (such as M-CSF or GM-CSF) on monocytic versus granulocytic lineage commitment of MDS cells are currently tested. To advance our study on the patient level we used primary high risk MDS CD34+ cells (N=36). Tumor cells with the lowest PU.1 expression displayed partial methylation at URE and incomplete sensitivity to AZA/G-CSF-mediated in vitro differentiation. We observed that patients with permissive chromatin at URE and higher PU.1 (and C/EBPa) levels responded significantly better to AZA/G-CSF in vitro. Currently we enrolled more patients in order to ask whether permissive PU.1 locus predicts responsiveness to AZA and whether it associates with different clinical patient outcome. (Grants: IGA NT10310-3/2009, MPO FR-TI2/509, GAUK251135 82210 & 251070 45410, SVV-2011-262507). Disclosures: No relevant conflicts of interest to declare.