Wenqi Diao, Wassim W Labaki, 2 MeiLan K Han, Larisa Yeomans, 3 Yihan Sun, 4 Zyad Smiley, 4 Jae Hyun Kim, Cora McHugh, 4 Pingchao Xiang, Ning Shen, Xiaoyan Sun, Chenxia Guo, Ming Lu, Theodore J Standiford, Bei He,* Kathleen A Stringer * 1Department of Respiratory Medicine, Peking University Third Hospital, Beijing, People’s Republic of China; 2Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, School of Medicine, University of Michigan, Ann Arbor, MI, USA; 3Biochemical Nuclear Magnetic Resonance Core, College of Pharmacy, University of Michigan, Ann Arbor, MI, USA; 4NMR Metabolomics Laboratory, Department of Clinical Pharmacy, College of Pharmacy, University of Michigan, Ann Arbor, MI, USA; 5Department of Respiratory and Critical Care Medicine, Shou-Gang Hospital Affiliated to Peking University, Beijing, People’s Republic of China; 6Department of Respiratory Medicine, Peking University Third Hospital, Beijing 100191, People’s Republic of China; 7Department of Respiratory Medicine, Peking University Health Sciences Center, Third Hospital, Beijing, People’s Republic of China

Background Chronic obstructive pulmonary disease (COPD) is a systemic condition that is too complex to be assessed by lung function alone. Metabolomics has the potential to help understand the mechanistic underpinnings that contribute to COPD pathogenesis. Since blood metabolomics may be affected by sex and body mass index (BMI), the aim of this study was to determine the metabolomic variability in male smokers with and without COPD who have a narrow BMI range. Methods We compared the quantitative proton nuclear magnetic resonance acquired serum metabolomics of a male Chinese Han population of non-smokers without COPD, and smokers with and without COPD. We also assessed the impact of smoking status on metabolite concentrations and the associations between metabolite concentrations and inflammatory markers such as serum interleukin-6 and histamine, and blood cell differential (%). Metabolomics data were log-transformed and auto-scaled for parametric statistical analysis. Mean normalized metabolite concentration values and continuous demographic variables were compared by Student’s t-test with Welch correction or ANOVA with post-hoc Tukey’s test, as applicable; t-test p-values for metabolomics data were corrected for false discovery rate (FDR). A Pearson association matrix was built to evaluate the relationship between metabolite concentrations, clinical parameters and markers of inflammation. Results Twenty-eight metabolites were identified and quantified. Creatine, glycine, histidine, and threonine concentrations were reduced in COPD patients compared to non-COPD smokers (FDR ≤15%). Concentrations of these metabolites were inversely correlated with interleukin-6 levels. COPD patients had overall dampening of metabolite concentrations including energy-related metabolic pathways such as creatine metabolism. They also had higher histamine levels and percent basophils compared to smokers without COPD. Conclusion COPD is associated with alterations in the serum metabolome, including a disruption in the histidine-histamine and creatine metabolic pathways. These findings support the use of metabolomics to understand the pathogenic mechanisms involved in COPD. Trial registration www.clinicaltrials.gov, NCT03310177.

Background: Though blood is an excellent biofluid for metabolomics, proteins and lipids present in blood can interfere with 1d-1H NMR spectra and disrupt quantification of metabolites. Here, we present effective macromolecule removal strategies for serum and whole blood (WB) samples. Methods: A variety of macromolecule removal strategies were compared in both WB and serum, along with tests of ultrafiltration alone and in combination with precipitation methods. Results: In healthy human serum, methanol:chloroform:water extraction with ultrafiltration was compared to methanol precipitation with and without ultrafiltration. Methods were tested in healthy pooled human serum, and in serum from patients with sepsis. Effects of long-term storage at −80 °C were tested to explore the impact of macromolecule removal strategy on serum from different conditions. In WB a variety of extraction strategies were tested in two types of WB (from pigs and baboons) to examine the impact of macromolecule removal strategies on different samples. Conclusions: In healthy human serum methanol precipitation of serum with ultrafiltration was superior, but was similar in recovery and variance to methanol:chloroform:water extraction with ultrafiltration in pooled serum from patients with sepsis. In WB, high quality, quantifiable spectra were obtained with the use of a methanol: chloroform precipitation.