BACKGROUNDSeveral biological mechanisms sustain leukemic development, but the inhibition of apoptotic death, which should be induced by the genetic instability of the disease, is surely one the key neoplastic alterations. Apoptotic regulation involves complex pathways of positive and negative regulators which may and do represent a target for personalized therapy. Furthermore, the overexpression of specific anti-apoptotic genes likely represent a useful prognostic instrument. Nowadays, BCL-2 inhibitors have successfully passed preliminary phases of clinical experimentation, however the involvement of other essential genes, such as MCL-1 and BCL2L1 (coding for bcl-xL), has been proposed as resistance mechanism to BCL-2 inhibitors and as novel therapeutic target in Acute Myeloid Leukemia (AML).Our study aims to investigate the expression of anti-apoptotic genes in AML patients to cluster them in different categories and study their impact on prognosis, in order to reveal possible mechanism of resistance to BCL-2 inhibitors and stratify patients that are more likely to take advantage of administration or co-administration of one between MCL-1 and BCL-2/BCL2L2 inhibitors. METHODSWe performed Human Transcriptome Array 2.0 (Affymetrix) in a cohort of 59 new diagnosed AML. Gene Expression Profiling (GEP) analysis has been conducted with Transcriptome Array V3.0 Software (Affymetrix). K-means clustering has been used to categorize different cluster of patients basing on BCL-2, MCL-1 and BCL2L1 gene expression. A pool of 7 healthy donor was used to normalize expression data. Survival analysis was conducted with Kaplan-Meyer method and differences in survival were assessed with Log-Rank test. RESULTSAnalyzing BCL-2, MCL-1 and BCL2L1 gene expression values in 59 AML patients at diagnosis we established 5 categories where patients are almost uniformly assigned (Figure 1a): 11/59 patients in Cluster 1, characterized by the only one overexpression of MCL1; 27/59 patients in Cluster 2, which presents low expression of all analyzed genes; 2/59 patients in Cluster 3, where BCL2L1 is strongly overexpressed; 6/59 patients in Cluster 4, distinguished by BCL-2 overexpression and 13/59 patients in Cluster 5, which high values of expression of all 3 genes. Thus, Clusters 2 and 5, where no gene is overexpressed in comparison with the others, include the majority of patients (40/59, approximately 68%), suggesting that apoptotic death control is realized thanks to an intricate debate between these 3 considered proteins and BCL-2 should not be considered the unique oncogenic anti-apoptotic mechanism. Furthermore, the 3-D Scatter Plot (Figure 1A) allows us to observe how the patients out of the central cloud (Clusters 2-5) always present one of the three anti-apoptotic genes overexpressed, and when BCL-2 level is progressively low (Cluster 1-2-3), MCL-1 and BCL2L1 values are high, respectively. Therefore, this documented behavior may indicate the presence of many ways used by leukemic cell to prevent apoptosis, with one of them almost always activated.In term of Overall Survival (OS), unsurprisingly no statistically significant difference arised between Cluster 1-3-4 where one different anti-apoptotic gene is overexpressed (Figure 1b), underling the probable equal effect of each mechanism in inhibiting apoptotic death and in sustaining leukemic progression. Finally, analyzing the impact only of BCL-2 expression on OS in a cohort of 36 young patients treated with chemotherapy it is possible to observe a trend close to statistically significance (p=0.07)(Figure 1c), unreached probably because of the relative limitation of the sample, confirming the BCL-2 induced chemoresistance. CONCLUSIONSApoptotic death regulation is precisely controlled equally by BCL-2, MCL-1 and BCL2L2, their expression is quite balanced in AML patients evaluated. One out of these three genes is almost always overexpressed, reaffirming the key role of anti-apoptotic oncogenes in leukemogenesis. Even though BCL-2 expression influence patients prognosis, MCL-1 and BCL2L2 overexpression has been documented when BCL-2 is normally or underexpressed. In conclusion, BCL-2 inhibitors resistance may be caused by MCL-1 and BCL2L2 overexpression detected in patients BCL-2 independents. Moreover, a synthetic lethality approach should be realized to switch-off leukemic anti-apoptotic multiple mechanisms. [Display omitted] DisclosuresMartinelli:Celgene: Consultancy; Amgen: Consultancy; Johnson&Johnson: Consultancy; Ariad/Incyte: Consultancy; Pfizer: Consultancy; Roche: Consultancy.