Nuclear magnetic resonance based metabolomic profiling in different types of leukemia

Leukemia is the most common cancer in childhood with a frequency of %30 and has a 3.2% incidence among all new cancer cases. Lymphoma in other respects differs in tissue origin but may manifest with similar symptoms to leukemia. Disease pathogenesis involves impairments that may belong to any maturation step of white blood cells in bone marrow/ lymphoid organs. Metabolomics literature harbors various works investigating whether defined biomarker molecules would be valuable for early diagnosis-prognosis or monitoring treatment response in leukemia/ lymphoma patients. In fact, these studies covering hundreds of metabolites offer great opportunities for identifying affected pathways related to the biomarkers and submit valuable information for individualized therapy options via multivariate statistical analysis. Nuclear magnetic resonance (NMR) and mass spectrometry (MS) constitute the two elemental instruments of metabolomics that precede each other in the measurement of different molecules because of their innate characteristics and so studies combining the two methods may cover the whole metabolite spectrum. NMR has several advantages as having high reproducibility, allowing real-time quantitation with only one internal standard for each run, long stability, and not requiring frequent calibration procedures. However, MS has still wider applications in the field with its high sensitivity allowing the detection of low-abundance molecules in mixtures, and being able to define a proteomic-lipidomic profile with high accuracy makes it the first choice in routine laboratories. The current dissertation study was designed to integrate the two methods of analyzing common types of leukemia and lymphoma with a global metabolomic profiling approach. In detail, serum and leukocyte isolate samples of acute myeloid leukemia, chronic lymphocytic leukemia, and non-Hodgkin's lymphoma patients (30 participant for each) were investigated on 1D, 2D 1H NMR and high-resolution MS platforms. For the metabolic profiling, quantitative 1H NMR (qNMR) and untargeted HR-MS methods were applied, and data were analyzed with univariate-multivariate statistics to identify differential metabolites and pathways (MetaboAnalyst, Tidymass). Metabolites identified in 1D experiments were confirmed by 2D NMR analyses. Leukocyte isolates were studied on the HRMS platform under equal conditions with serum samples, and metabolites that were significantly expressed compared to controls were determined (TidyMass). According to the serum NMR analysis; metabolites increased in the AML group compared to controls (VIP score > 1) were recorded as methylguanidine, isobutyrate, glycine, 2-hydroxyisobutyrate, glucose, urea, betaine, and those that decreased (VIP score > 2) were pyroglutamate, 3-hydroxyisovalerate, alanine. The metabolites increased in the CLL group (apart from common ones with AML) included creatinine-creatine phosphate, tyrosine and phosphoethanolamine molecules. The fact that the molecules increased in all groups are protein degradation products confirms the rapid protein turnover observed in hematological malignancies. The decrease in fumarate and acetate molecules with a VIP score greater than 2.5 also indicates the intense consumption of Krebs cycle intermediate components in proliferating leukoblasts, as common differences in the CLL and NHL groups. In addition, the low serum carnitine level (VIP score > 1), which is a specific difference in the NHL group, suggests that the increased energy need in lymphoma patients is compensated by an increase in beta oxidation. The receiver operating characteristic (ROC) analyzes including serum NMR results showed area under the curve values greater than 0.98 for diagnostic biomarker models with 10 top metabolites for AML, 3 for CLL, and 20 for NHL. In pathway analyzes showing group-based variation; glutathione, glucogenic/ branched-chain amino acid metabolism and choline oxidation ranked first in AML while Krebs and urea cycle were prominent in the CLL and NHL groups. According to the serum HRMS results, distinct polar lipids (fatty acids and phospholipids) were identified with the help of highly sensitive methodology, and the unsaturated fatty acids that increased significantly in the AML group were recorded as docosatrienoic with cis-13-docozenoic acid (p=0.002 and 0.003, respectively, Tidymass). Serum HRMS results also revealed that oleamide was significantly increased in the CLL group, and glycerophosphocholine and pyroglutamate were found to be at low levels, confirming the NMR results (p = 0.001 and 0.000, respectively, Tidymass). NHL serum HRMS findings were similar to AML in terms of increased unsaturated fatty acids and to CLL in terms of oleamide while decreased pyroglutamate levels were observed as CLL. In HRMS analyzes performed on leukocyte isolates, fewer significant metabolites were identified compared to serum as increased levels of spermine and long-chain fatty acids (supporting the serum results) with low levels of coenzymes such as biotin that were common in all groups. In conclusion, complementary use of NMR-MS methodologies expanding the metabolite spectrum in both matrixes allowed comprehensive profiling of leukemia patients. The current findings eliciting group-based changes provide essential tips about pathogenetic mechanisms commonly observed in distinct subclasses along with individual driving forces. In the perspective of personalized medicine, identifying custom targets and monitoring aberrant changes in affected individuals is of great importance and could only be possible by virtue of using strengths of metabolomics tools as NMR and MS. Hence, this work also contributed to the future construction of a workflow for hematological malignancy inspection to be used in routine hospital laboratories in the metabolomics boom era. Yet, emerging results need to be validated with wider cohorts and omics information for each patient within the frame of personalized medicine approach.