ALBERT

Description: Despite its clinical activity in myelodysplastic syndrome (MDS), the relationship between the inhibition of DNA methyltransferase 1 (DNMT) by 5-azacitidine (5AC) and hematologic improvement has not been well-explored. Optimal in vitro re-expression of genes silenced through promoter methylation requires sequential exposure to a DNMT inhibitor (i) and an histone deacetylase (HDAC)i. In a dose/schedule exploration study, patients with MDS (including MDS-AML) were treated with various regimens of 5AC SQ, followed by a 7 day infusion of the HDACi sodium phenylbutyrate (PB). Patients received 5AC at the following doses: 50 mg/m2/day * 10 doses (cohort A, n = 8), 50 mg/m2/day * 14 doses (cohort B, n = 3), and 25 mg/m2/day * 14 doses (cohort C, n = 5). Patients received a minimum of 4 cycles (q28 days). Clinical responses were graded according to IWG criteria. Changes in promoter methylation of p15INK4B and E-Cadherin (E-CAD), the most commonly methylated genes in MDS and AML, were monitored using a real time-PCR modification of methylation specific-PCR (rtMSP). Changes in gene expression were monitored using real time rtPCR. Dose-limiting hematologic toxicity (myelosuppression 〉 14 days) occurred in 2/3 patients in cohort B. The other two dose schedules were well-tolerated. Clinical responses developed in 4/8 patients in cohort A (3 CR, 1 PR), 2/3 patients in cohort B (1 hematologic improvement (HI)-P, N, major, 1 HI-P-major), and 3/3 currently evaluable patients in cohort C (2 HI-N-major, 1 HI-P, major). p15 was methylated in 10/10 evaluable samples pre-treatment; E-CAD was methylated in 5/7. Treatment with 5AC decreased p15 methylation in 50% of patients studied. p15 expression increased in patients in whom methylation was decreased; this included 3/4 clinical responders studied. In several cases, reversal of methylation was confirmed by bisulfite sequencing (BSQ) of serial samples. BSQ suggested that 5AC treatment led to gradual demethylation of the clone, rather than replacement of a methylated clone with a normal unmethylated clone. Surprisingly, treatment with 5AC increased global acetylation of histones H3 and/or H4 (Western analysis) in 7/8 patients in cohort A, and 2/2 evaluable patients in cohort B, (cohort C data pending). Acetylation was further increased following PB administration in 4/8 patients in cohort A and 1/2 in cohort B. These data confirm for the first time that clinical administration of 5AC leads to substantial reversal of promoter methylation associated with gene re-expression. Administration of 5AC is also associated with induction of global histone acetylation; the mechanism underlying histone acetylation in response to 5AC is unclear. While administration of 5AC and PB is associated with re-expression of p15, the relative contributions of the DNMT and HDAC inhibitors cannot be determined from the present study. Gene methylation data obtained using rtMSP correlates well with BSQ. The use of this semi-quantitative technique to monitor larger studies of 5AC with and without HDAC inhibitors will facilitate determination of the relationship between reversal of promoter methylation of p15 and other genes and clinical response to DNMTis.

Description: Death-associated protein kinase (DAP-Kinase) is a novel serine/threonine kinase whose expression is required for γ interferon-induced apoptosis. A previous study suggested that DAP-Kinase expression may be lost epigenetically in cancer cell lines, because treatment of several nonexpressing cell lines with 5-aza-2′-deoxycytidine resulted in the expression of DAP-Kinase. Using methylation-specific polymerase chain reaction (MSP), we examined the DAP-Kinase CpG island for hypermethylation in cancer. Normal lymphocytes and lymphoblastoid cell lines are unmethylated in the 5′ CpG island of DAP-Kinase. However, in primary tumor samples, all Burkitt’s lymphomas and 84% of the B-cell non-Hodgkin’s lymphomas were hypermethylated in the DAP-Kinase CpG island. In contrast, none of the T-cell non-Hodgkin’s lymphoma samples and 15% or less of leukemia samples examined had hypermethylated DAP-Kinase alleles. U937, an unmethylated, DAP-Kinase–expressing leukemia cell line, was treated with γ interferon and underwent apoptosis; however, Raji, a fully methylated, DAP-Kinase nonexpressing Burkitt’s lymphoma cell line, only did so when treated with 5-aza-2′-deoxycytidine followed by γ interferon. Our findings in cell lines and primary tumors suggest that hypermethylation of the DAP-Kinase gene and loss of γ interferon-mediated apoptosis may be important in the development of B-cell malignancies and may provide a promising biomarker for B-cell–lineage lymphomas.

Description: Several lines of evidence suggest that genomic instability in myeloid malignancies is promoted by increased endogenous DNA damage and error-prone repair that lead to disease progression and resistance to therapy. We recently reported increased levels of Poly-(ADP)-ribose polymerase (PARP) and DNA Ligase IIIα as well as increased activity of a highly error-prone pathway for repair of DNA double-strand breaks in myeloid leukemias. Importantly, these leukemia cells are sensitive to inhibitors of DNA Ligase IIIα and PARP, suggesting their dependence on these factors for survival. Parallel studies have shown that transient exposure to DNA demethylating agents at low nM concentrations reprograms cancer cells, altering heritable gene expression patterns in key cellular pathways, including DNA repair pathways, suggesting that pre-treatment of leukemia cells with demethylating agents may further sensitize them to PARP inhibitors. Thus, established cell lines from acute myeloid leukemia (AML; MV411, KASUMI-1), myelodysplastic syndrome transformed to AML (MDS; P39) and bone marrow mononuclear cells obtained from AML patient samples (N=7) were exposed to non-cytotoxic doses of DNA methyltransferase inhibitors (DNMTis, decitabine, DAC), followed by four days without drug treatment and subsequent treatment with low doses of PARP inhibitors (PARPis; ABT888, or BMN673) alone or in combination with DNMTis. Clonogenicity, apoptosis, DNA repair efficiency, and activity and expression level of DNA repair and DNA methyltransferase proteins were then studied. In all the cell lines tested, treatment with DAC (5-10nM) followed by ABT888 (500nM) induced a significant decrease in colony survival compared to control or single treatment. The use of a more potent PARPi, BMN673 (0.1-10nM), confirmed that treatment with DNMTis followed by PARPis induces a robust inhibition of AML and MDS cell line colony forming capacity. Interestingly, the same schedule treatment of decitabine followed by PARPis significantly decreases the clonogenic capacity in 4 out of 7 (57%) of bone marrow mononuclear cells from AML patient tested so far, suggesting that DNMTis and PARPis sequential treatment could be a valuable therapeutic option for AML and MDS patient. We next initiated studies to elucidate the mechanism by which DAC may sensitize myeloid malignancies to PARPis. As expected, DAC treatment alone was sufficient to decrease DNMT1 expression levels and increase caspase 3 cleavage in AML cell lines, compared to control treated cells. But surprisingly, DAC treatment alone also induced a decrease in PARP protein expression, with a further decrease in cells treated with DAC followed by PARPis, suggesting that both methylation and DNA repair signaling alter PARP1 steady-state levels. Moreover, preliminary results show that the presence of PARP on chromatin is decreased with DAC treatment and further decreased following PARPis. In conclusion, our results suggest that DNMTis reprogram cells, sensitizing them to PARP inhibition in AML/MDS patient and cell line models, paving the way for testing the therapeutic potential of sequential treatment with these agents in clinical trials. We are exploring one hypothesis that decreased levels of PARP on chromatin following DAC treatment may lead to more effective trapping of PARP1 at sites of DNA damage by PARPis, leading to abrogation of DNA repair. Understanding how these proteins interact may explain the mechanisms underlying the sensitization of epigenetically reprogrammed cells to PARPis, and may define the molecular subsets of AML patients that may respond to this novel therapeutic strategy. Disclosures: No relevant conflicts of interest to declare.

Description: The promoter region of the cyclin-dependent kinase inhibitorp15INK4Bcontains a CpG island that is hypermethylated in many hematologic malignancies. To explore the relationship between patterns of methylation and gene transcription, we used bisulfite genomic sequencing to obtain a detailed analysis of methylation in acute leukemia, leukemia cell lines, and normal lymphocytes. The entire CpG island region of p15 was largely devoid of methylation in normal lymphocytes, but methylation of varying density was found in primary acute leukemia. Methylation density was generally conserved between the alleles from each sample, but marked heterogeneity for the specific CpG sites methylated was observed. Patterns of methylation were compared and expression assessed with reverse-transcriptase polymerase chain reaction (RT-PCR). The density of methylation within the CpG island, and not any specific location, correlates best with transcriptional loss. Leukemias with methylation of approximately 40% of the CpG dinucleotides on each allele had complete gene silencing, with variable, but diminished expression with less dense CpG island methylation. Our results suggest that the transcriptional silencing of p15 in conjunction with aberrant hypermethylation is best understood as an evolutionary process that involves progressively increasing methylation of the entire p15CpG island.

Description: Abstract SCI-28 The past decade has seen an explosion in our knowledge about epigenetic regulation of gene expression in both normal and neoplastic cell renewal. We can increasingly view cancer, including lymphatic malignancies, as diseases driven by both genetic and epigenetic abnormalities. Cancer epigenetics began with focus on alterations in DNA methylation leading to our current knowledge that every individual patient's tumor harbors hundreds of abnormally, transcriptionally silenced, genes which harbor CpG island hypermethylation in their promoters. Now, we understand that what might be termed the “cancer epigenome” is constituted by many other chromatin aberrations and focus now includes alterations in histone acetylation and other histone modifications, polycomb group protein (PcG) transcriptional repression, and interaction of each of these processes with DNA methylation changes. Additionally, we are increasing our understanding of the molecular processes which initiate and maintain epigenetic abnormalities in cancer by realizing their ties, in multiple ways, to links between chromatin regulation in embryonic stem (ES)/progenitor cells and cancer stem like cells (CSC's). First, the genes affected with abnormal silencing in cancer include those which affect virtually every key cell signaling pathway common to maintenance of stemness and tumorigenesis including control of the cell cycle and of stem/progenitor cell regulation (p16, p15), the Wnt pathway (APC, SOX17), cell adhesion and invasion (E-cadherin), and apoptosis (DAP-kinase). Second, there is an emerging concept of a molecular progression of gene silencing during tumorigenesis. This begins with a link to a feature of the embryonic epigenome which is normally responsible for maintaining key ES genes in a low, but poised, transcription state. When associated with cancer, this state may contribute to disruption of normal lineage commitment and differentiation and contribute to abnormal cell expansion in early stages of cancer evolution and to the phenotype of CSC's. The core of this molecular progression involves initial PcG repression of transcription of genes with promoter CpG islands not normally DNA methylated during development. Abnormal DNA methylation is subsequently imposed upon these promoters in cancer deepening their transcriptional repression. All of these tumor changes may be spurred by cancer risk states including ageing and abnormal cell renewal during chronic inflammation. Multiple translational implications have arisen based on our increasing understanding of the cancer epigenome. First, abnormal gene promoter DNA methylation, and other chromatin changes, can be utilized as potentially powerful biomarkers for cancer risk assessment, early cancer diagnosis, molecular staging of tumors, and monitoring of drug sensitivities. Second, there is current excitement about the promise of “epigenetic therapy” centering around “reprogramming” the cancer epigenome as a molecular target strategy. This has potential as frontline cancer therapy and for treatment of advanced disease. Importantly, cancer epigenetic control of CSC's has been recently linked to maintenance of their drug resistant phenotypes and to the possibility of reversing this resistance by epigenetic therapies. Disclosures: Baylin: Oncomethylome Sciences: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; BioNumerick Pharm: Consultancy, Research Funding; Constellation Pharm: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Progen: Consultancy, Membership on an entity's Board of Directors or advisory committees.

Description: Death-associated protein kinase (DAP-Kinase) is a novel serine/threonine kinase whose expression is required for γ interferon-induced apoptosis. A previous study suggested that DAP-Kinase expression may be lost epigenetically in cancer cell lines, because treatment of several nonexpressing cell lines with 5-aza-2′-deoxycytidine resulted in the expression of DAP-Kinase. Using methylation-specific polymerase chain reaction (MSP), we examined the DAP-Kinase CpG island for hypermethylation in cancer. Normal lymphocytes and lymphoblastoid cell lines are unmethylated in the 5′ CpG island of DAP-Kinase. However, in primary tumor samples, all Burkitt’s lymphomas and 84% of the B-cell non-Hodgkin’s lymphomas were hypermethylated in the DAP-Kinase CpG island. In contrast, none of the T-cell non-Hodgkin’s lymphoma samples and 15% or less of leukemia samples examined had hypermethylated DAP-Kinase alleles. U937, an unmethylated, DAP-Kinase–expressing leukemia cell line, was treated with γ interferon and underwent apoptosis; however, Raji, a fully methylated, DAP-Kinase nonexpressing Burkitt’s lymphoma cell line, only did so when treated with 5-aza-2′-deoxycytidine followed by γ interferon. Our findings in cell lines and primary tumors suggest that hypermethylation of the DAP-Kinase gene and loss of γ interferon-mediated apoptosis may be important in the development of B-cell malignancies and may provide a promising biomarker for B-cell–lineage lymphomas.