Investigations of Serum Metabolomic Relationships between Lung Cancer and Alzheimer’s Disease using Nuclear Magnetic Resonance Spectroscopy

Abstract

AbstractBackgroundAlzheimer’s Disease (AD) and Lung cancer (LuCa) are among the leading causes of mortality in the United States. Remarkably, several studies demonstrated a potential inverse co‐morbidity of AD with LuCa. Possible explanations related to the cell cycle (dys)regulation include presence of specific neuropathological features, immune checkpoint functions and increased circular RNA levels. Metabolomic profiles for LuCa or AD alone have been proposed. Here, using high‐resolution magic angle spinning (HRMAS) MRS, we aim to study metabolomic relationship between the two diseases with serum metabolomic profiles for matched LuCa patients, with and without AD, by comparing serum profiles measured from matched healthy controls.MethodsLuCa sera from of patients with (LCAD, n = 10, 3M, 7F, 77.3±7.7yrs) and without (LCnonAD, n = 10, 3M, 7F, 77.4±5.2yrs) clinically defined AD were obtained from Harvard/MGH Lung Cancer Susceptibility Study (a.k.a. BLCS) Repository, with individually age‐ and gender‐matched healthy controls (Ctrl‐LCAD and Ctrl‐LCnonAD, n = 10/each) obtained from the MGB BioBank. HRMAS proton NMR for these 40 samples were measured on 10 µl, at 4°C with a spin rate of 5000 Hz using a Bruker Avance 600 MHz spectrometer. Spectra were analyzed with Bruker Topspin 3.6.2. Statistical analysis was performed using JMP Pro on 18 spectral regions identified from HRMAS proton NMR.ResultsAs a survey of the entire data ensemble of 18 spectral regions measured from 40 samples, the hierarchical clustering analysis (Fig. 1) seems to associate different groups into different clusters. Unsupervised principal component analysis (PCA) conducted on all spectral regions generate PCs (PC1 and PC2) that can differentiate among groups (Fig. 2). Further noticed that, four out of 18 regions can significantly differentiate LCAD from LCnonAD. Calculations of PCA using these four regions result in PC1_4regions that can significantly differentiate LCAD from LCnonAD (Fig. 3). Collectively evaluating potential contributing metabolites towards PC1_4regions indicate increases in myo‐inositol suggesting towards LCAD, which increases in creatine, phosphocreatine, serine, aspartate, tyrosine, proline, glutamine, and 2‐oxoglutarate suggesting towards LCnonAD.ConclusionOur findings, whilst preliminary, seem promising in guiding further investigations into serum NMR metabolomics for better understanding LuCa‐AD metabolic relationships.