Initial Analysis of Lipid Metabolomic Profile Reveals Differential Expression Features in Myeloid Malignancies

Introduction: Lipids are molecules that stand out among the different cellular metabolites by their enormous molecular diversity. Their functions were initially related to the composition of biological membranes and energy storage, but currently, these molecules have been analyzed considering different functions and regulatory signaling (Loizides-Mangold, 2013). It is known that lipid membranes and lipid mediators constitute specific phenotypes, including tumors (Hilvo et al, 2011). Lipid metabolism in cancer had been studied predominantly at the genetic level has recently gained further interest. Lipidomics studies show a powerful means of investigating pathophysiological issues and involvement of lipids in pathological states for both diseases in which lipids are known to play a role, but also for those which role is not well characterized (Roberts et al., 2008). Myeloid neoplasms are clonal diseases of the hematopoietic stem cell which can be present in the bone marrow and/or peripheral blood. Over the past two decades, mass spectrometry (MS) has emerged as the main method used in lipidomics analysis, which allows the structural characterization and quantification of complex lipids and their metabolites (MURPHY et al, 2005). Due to the importance of this field we have considered the use of the lipidomic innovative platform to identify differences in the plasma lipid metabolomic profile of hematological patients with Myeloid Neoplasms.

Methods: Untargeted Shotgun MS/MS Analysis was performed on an independent service at the AB-Sciex Laboratory located in Sao Paulo, SP, Brazil on a 5600 Triple TOF mass spectrometer (ABSciex) instrument with an acquisition scan rate of 100 spectra/sec and stable mass accuracy of ~2 ppm. Plasma samples from 153 participants were analyzed being, 90 of the Control Group, 43 Myeloproliferative Neoplasms (MPN), 11 Myelodysplastic Syndromes (MDS) and 9 Acute Myeloid Leukemias (AML). Data were acquired using the AB-Sciex Analyst TF, processed using the AB-Sciex LipidViewTM and the web-based analytical pipeline MetaboAnalyst 2.0 (www.metaboanalyst.ca) (Xia et al, 2012).

Results: Untargeted analysis identified in negative and positive-modes a total of 658 features at 2 ppm resolution. PCA and PLS-DA analysis revealed clear discrimination among groups, in particular for AML patients. Main lipid groups differentially expressed were: Monoacylglycerols (MAG), Glucosylceramide E (GlcdE), Ethyl Esters (EE), Lysophosphatidic acid (LPA), Sulfoquinovosil diacylglycerols (SQDG), Monoglycerols (MG), Methyl Ethanolamines (ME), Lysophosphatidylcholines (LPC), Dimethyl Phosfatidyletanilamines (DMPE), Monometylphosphatidiletanolamines (MMPE), Ceramide-1-phosphate (CerP), Glicerophosphoglycerols (GP), Lysomonomethyl Glycerophosphocholine (LMMPE), Phosphatidic Acids (PA), Ergosterols (ERG), Glycerophosphoserine (PS), Diacylglycerols (DAG), Hexocylceramides (HexCer) and Lanosterol (Lan). ROC Curve Analysis revealed Total LMMPE as the strongest discriminating marker between Controls from Patients with MDS or AML (Sensitivity= 0.95 (0.824-1); Specificity= 0.8941 (0.847-0953); Positive Likelihood Ratio= 8.972 and Negative Likelihood Ratio =0.05592 and T Test= 7.576E-12). In addition these lipids were also able to differentiate MDS and AML from NMP (Sensitivity= 0.9118 (0.824-1), Specificity= 0.95 (0.85-1), Positive Likelihood Ratio= 18.2 and Negative Likelihood Ratio= 0.05592).

Conclusions: The Myeloproliferative Neoplasms from the point of view of global plasma lipidomics are accompanied by several modifications. In particular the Lysomonomethyl-Phosphatidylethanolamines (LMMPE) seems to play important differentiating roles among them.