Abstract 2790 Background:TKIs were initially developed as targeted therapies that would solely interfere with a “specific” aberrant signaling pathway in malignant cells. However, we and others showed that the EGFR TKI erlotinib (Erlo) has in vivo and in vitro efficacy in MDS and AML (Boehrer et al., Blood, 2008). We also previously observed, in a preliminary study, the potentiation of apoptosis upon combination of Erlo with azacitidine (Aza), but not with decitabine, in HL-60 cells and in cells from a few (five) MDS/AML patients (ASH 2010, 974). We decided to expand this pre-clinical study to define the potential interest of combining different TKI (Dasatinib (Dasa), Sorafenib (Sora) or Erlo) with Aza, now a reference first line treatment in higher risk MDS (Lancet Oncol, 2009). Methods:Erlo (10μM), Dasa (500nM), or Sora (5μM) were combined to Aza (1μM) and apoptosis over-time (24, 48 and 72h) quantified by FACS analysis following DioC3(6)/PI staining in different MDS/AML-derived cell lines (MOLM-13, SKM-1, MV4-11, Kasumi-1). Quantification of apoptosis at 48h or 72h was recapitulated ex vivo in CD34+ cells from patients with MDS (n=12), AML (n=14) or AML post MDS (n=5). For each single drug, as well as the respective combinationsErlo, Aza and Erlo+Aza, the capacity to induce cell cycle arrest (PI staining), cytotoxicity (MTT assay) and decrease of proliferation (Click-it EdU Assay) was assessed concomitantly. Differentiation was assessed by staining with CD11b on day 3 (FACS), evaluation of ROS production at 24/48h by FACS staining with CM-DFCDA (oxidative stress indicator) and HE (Hydroethidine). Functional relevance of apoptosis-related signaling pathways was determined by co-incubation of Aza (and drugs combinations with biochemical inhibitors against MAPK: JNK (SP600125, 10μM), p-38MAPK (SB203580, 10μM) and MEK (U0126, 5μM). Immunoblot analyses of caspase-3, PARP, Mcl-1 and Bcl-xl were performed at 24h and 48h. Results:Whereas co-incubation of Dasa or Sora with Aza did not increase the degree of apoptosis observed with Aza alone, combination of Erlo with Aza had synergistic effects already observed at 24h, an effect that increased over-time (SKM-1 72h (mean PI+ cells, n=3), Erlo: 15%, Aza: 40%, Erlo+Aza: 81% - MOLM-13 72h, Erlo: 21%, Aza: 25%, Erlo+Aza: 64%). The % of cells with intact metabolic activity at 72h (determined by the MTT assay using 2.5μM Erlo and 0.5μM Aza) decreased from 83% (Erlo) and 79% (Aza) to 23% (Erlo+Aza) in SKM-1 (similar results for MOLM-13 cells). To determine if Erlo also impacts on apoptosis in CD34+ cells from patients with MDS or AML, we screened 31 samples and observed a synergistic effect in 5/31 samples (2/12 MDS, 2/14 LAM, 1/5 AML post MDS) and an additive effect in 8/31 samples. Induction of apoptosis was not preceded by differentiation (no increase in CD11b) or production of ROS. On the other hand, analysis of cell cycle distribution at 24h showed that apoptosis was accompanied by an increase in G0/G1 arrest and a decrease in the % of cells in S phase upon incubation with the combination of Erlo+Aza (G0/G1 (mean, n=2) in SKM-1, DMSO: 51%, Erlo: 52%, Aza: 61%, Erlo+Aza: 78%, G0/G1 in MOLM-13, DMSO: 40%, Erlo: 51%, Aza: 38%, Erlo+Aza: 65%) confirmed by the study of newly synthesized DNA before induction of apoptosis (15h) with an alternative BrdU test (Click-EdU assay). Indeed, in SKM-1 cells, proliferation was decreased by 20% upon Erlo or Aza and by 80% with the association. Moreover, cell cycle arrest was associated with an increase in the % of apoptotic cells with a sub-diploid DNA content (about 10% for Erlo or Aza and 32% for Erlo+Aza in SKM-1. Immunoblot analyses confirmed an increase in cleaved PARP and caspase-3 upon incubation with Aza +Erlo as compared to the single agents, as well as a decrease of the anti-apoptotic protein Mcl-1 and Bcl-xl. Incubation of Aza with SP600125 induced the same extent of apoptosis observed with Erlo+Aza, whereas no effect was observed with the p-38MAPK inhibitor SB203580 or the MEK inhibitor U0126. In addition, Erlo decreased phosphorylation of JNK on Thr183/Tyr185, suggesting a potential implication of the JNK pathway in the mechanism of apoptosis. Conclusions:In this study, Erlo was the only TKI tested able to increase sensitivity towards Aza in MDS/AML cell lines and in patient-derived cells. These results suggest a potential clinical interest of combining Aza to Erlo in MDS/AML. Disclosures:Fenaux:Celgene: Honoraria, Research Funding.