Identification of Molecular Targets of AML by Phosphoproteomic Screening of Valproic Acid Treated BNML and C. Elegans RNAi Validation.

Abstract 4151

Disease stabilisation, instead of cure, is proposed as the therapeutic strategy of choice in elderly or chemoresistant acute myeloid leukemia (AML). This approach may also be of particular benefit to patients for whom allogeneic bone marrow (re)transplantation is not an option. Previously, we have clinically investigated the addition of valproic acid (VPA) to various combination chemotherapies with initial results in AML indicating prolonged survival, with follow-up on a predominantly out-patient basis. Consequently, we aimed to identify further molecular targets of VPA, which may enhance its therapeutic efficacy through screening for VPA-modulated phosphoproteins in relevant preclinical models of AML, and validation of these targets in RNAi screen of Caenorhabditis elegans (C. elegans).

Brown Norwegian Myeloid Leukemia (BNML) has previously been described as a particularly relevant preclinical rat model of AML. Indeed, leukemic rats treated with 170 mg/kg VPA twice-daily achieved therapeutic serum levels of VPA and demonstrated significant increases in survival in comparison to controls (p = 0.004). To screen for molecular targets of VPA effect in this responsive model, we investigated the differences in control and VPA treated BNML phosphoproteomes by difference gel electrophoresis (DIGE) separation and subsequent differential gel software analysis. This was achieved through harvest of phosphoproteins from leukemic blasts, isolated from the spleens of treated and control BNML rats by immobilized metal ion affinity chromatography (IMAC) and subsequent protein identification via Orbitrap mass-spectrometry. Significant differential expression of 9 phosphoproteins was found in VPA treated BNML rats compared to controls, including Tubulin α-1B chain (TBA1B) and Actin β (ACTB), indicating these genes as possible targets of VPA therapy.

To validate the functionality of 7 of these genes, RNAi was performed in wild type Bristol N2 strain of C. elegans at larval stage L1, 24 hours prior to exposure to 15 mM VPA for 72 hours. Knockdown of 4 of 7 genes resulted in larval developmental arrest, defined as synthetic lethality. In order to ascertain if synthetic lethality induced by these 4 genes was resultant of apoptosis, we employed the CED-1::GFP transgenic reporter assay to quantify germline cell death following RNAi depletion and subsequent exposure to VPA (15 mM, 24 hours). Increased numbers of apoptotic corpses in the germline was determined for all genes examined. To further examine the role of p53 in the observed apoptotic induction we used the transgenic strain cep-1::CED-1::GFP, which expresses the C. elegans ortholog of p53, CEP-1. Successive RNAi knockdown of our 4 candidate genes, again effected increased basal number of apoptotic corpses independently of CEP-1. These results suggest that similar combinational treatment of AML may be beneficial, irrespective of p53 status.

To further investigate this thesis in a human AML cell line, MOLM-13 cells were co-treated with VPA and small molecule inhibitors of prospective targets TBA1B and ACTB, namely paclitaxel, and cytochalasin B. Inhibition of actin polymerization or stabilisation of tubulin polymerization resulted in increased apoptosis when supplemented with VPA, as determined by DNA specific staining with Hoechst 33342. These results suggest that use of these combinations may be beneficial in the treatment of AML. In conclusion, this study indicates that phosphoproteomic screening of BNML and subsequent target verification in C. elegans worms has the potential to identify future drugable targets for effective combinatorial therapy with valproic acid in acute myeloid leukemia.