J Cancer Res Clin Oncol DOI 10.1007/s00432-017-2347-0 ORIGINAL ARTICLE – CANCER RESEARCH Study of early stage non-small-cell lung cancer using Orbitrap- based global serum metabolomics Agnieszka Klupczynska1 · Paweł Dereziński1 · Timothy J. Garrett2 · Vanessa Y. Rubio2 · Wojciech Dyszkiewicz3 · Mariusz Kasprzyk3 · Zenon J. Kokot1 Received: 10 November 2016 / Accepted: 18 January 2017 © The Author(s) 2017. This article is published with open access at Springerlink.com Abstract a number of NSCLC biomarker candidates which belong Purpose The aim of the project was to apply ultra-high- to such compound classes as acylcarnitines, organic acids, performance liquid chromatography–quadrupole-Orbitrap- and amino acids. Multivariate ROC curve built using 12 high-resolution mass spectrometry for serum metabolite identified metabolites was characterized by AUC = 0.836 profiling of non-small-cell lung cancer (NSCLC). This (0.722–0.946). There were no significant differences in the Orbitrap-based methodology has been applied for a study serum metabolite profiles between two most common his- of NSCLC potential markers for the first time. tological types of lung cancer—adenocarcinoma and squa- Methods After extraction using protein precipita- mous cell carcinoma. tion, sera were separated on the ACE Excel 2 C18-PFP Conclusions Through identification of novel potential (100 × 2.1 mm, 2.0 µm) column using gradient elution and tumor markers, Orbitrap-based global metabolic profiling analyzed within the range of 70–1000 m/z. Only patients is a useful strategy in cancer research. Our study can accel- with early stage disease (stages IA–IIB) were included in erate development of new diagnostic and therapeutic strate- the study, providing opportunity to find biomarkers for gies in NSCLC. The metabolites involved in discrimination early lung cancer detection. The resulting metabolite pro- between NSCLC patients and the control subjects should files were subjected to univariate and multivariate statisti- be further explored using a targeted approach. cal tests. Results 36 features were found significantly changed Keywords Lung cancer · Global metabolomics · between NSCLC group and controls after FDR adjustment Metabolite profiling · Mass spectrometry · Orbitrap and 19 were identified using various metabolite databases (in-house library, HMDB, mzCloud). The study revealed Introduction Electronic supplementary material The online version of this article (doi:10.1007/s00432-017-2347-0) contains supplementary Recently, the growing trend of metabolomic studies of can- material, which is available to authorized users. cer has been observed. It is well known that cancer cells exhibit an altered metabolism and increased energy require- * Zenon J. Kokot [email protected] ments compared to normal cells. However, the tumor metabolome has not yet been fully explained (Aboud and 1 Department of Inorganic and Analytical Chemistry, Poznan Weiss 2013). In contrast to normal cells, which utilize oxi- University of Medical Sciences, Grunwaldzka 6 Street, dative phosphorylation for energy production, cancer cells 60-780 Poznan, Poland 2 gain energy from glycolysis occurring even in the presence Department of Pathology, Immunology and Laboratory of oxygen. This phenomenon is called the Warburg effect Medicine, College of Medicine, University of Florida, 1395 Center Drive, Gainesville, FL 32610, USA (Vander Heiden et al. 2009). Another distinctive trait of cancer cell metabolism is the use of glutamine as a main 3 Department of Thoracic Surgery, Poznan University of Medical Sciences, Szamarzewskiego 62 Street, source of energy (Reitzer et al. 1979; DeBerardinis et al. 60-569 Poznan, Poland 2008). The characteristic metabolic alterations occurring Vol.:(0123456789)1 3 J Cancer Res Clin Oncol in tumors might result in changed levels of some metabo- are still several gaps and inconsistencies in the knowledge lites in tissues and body fluids. Human biological fluids are about the serum metabolome of lung cancer patients. commonly recognized as vehicles for the transmission of Only detection of early metabolic alterations can lead markers of many human disorders and thus the metabolite to early lung cancer diagnosis. Therefore, it is of particu- profiling of different body fluids is a promising strategy in lar interest to study the utility of cancer-related metabolic cancer research. There is a multitude of potential applica- changes occurring at the beginning of the disease. Many of tions of metabolomics in oncology, including early and the metabolomic studies included late-stage cancer cases accurate diagnosis, estimation of treatment efficacy, and (Hori et al. 2011; Fahrmann et al. 2015; Miyamoto et al. development of novel anti-cancer therapies (Spratlin et al. 2015). Therefore, the putative biomarker metabolites iden- 2009). tified in those studies would likely fail a further valida- Application of high-throughput and sensitive analyti- tion step for early detection of cancer. It should be noted cal techniques that are used in metabolomics makes them that by including advanced stage patients in the research, powerful tools in the field of oncology and aids understand- the obtained results could be overoptimistic. Moreover, ing what is happening in cancer cells. Given the complex there is a risk, especially in the case of amino acids, that nature of cancer, global metabolomic profiling offers the the observed abnormalities are influenced by nutritional unique opportunity of broadening the knowledge on the deficiencies and accompanying weight loss experienced tumor metabolome. In particular, there is an increasing by many patients with advanced lung cancer (del Ferraro interest in the study of metabolic changes related to lung et al. 2012). The application of metabolomic research in cancer over the last few years (Hori et al. 2011; Chen et al. the discovery phase for new cancer biomarkers is indisput- 2014; Deja et al. 2014; Puchades-Carrasco et al. 2016). able; however, the selection of representative samples has a Lung cancer constitutes one of the greatest challenges major impact on any conclusions drawn from the research. in contemporary oncology because of difficulties in early The aim of the study was to apply an Orbitrap-based detection resulting in lung cancer as a leading cause of global metabolomic approach to the analysis of serum of cancer death for many years (Subramaniam et al. 2013). patients with non-small-cell lung cancer (NSCLC). To Blood metabolite profiling of lung cancer includes the our knowledge, this is the first study which presents the application of such analytical techniques as gas chroma- application of ultra-high-performance liquid chromatogra- tography–mass spectrometry (Hori et al. 2011; Musharraf phy–quadrupole-Orbitrap-high-resolution mass spectrom- et al. 2015; Fahrmann et al. 2015), ultra-high-performance etry (UHPLC-Q-Orbitrap-HRMS) for searching NSCLC liquid chromatography–linear ion trap-mass spectrometry potential markers. Until now, Orbitrap-based profiling has (Mazzone et al. 2016), high-performance liquid chromatog- only been used to compare two metabolomes (plasma and raphy–quadrupole-time-of-flight-mass spectrometry (Chen serum) of small-cell lung cancer (SCLC) patients undergo- et al. 2014; Li et al. 2014), and nuclear magnetic resonance ing treatment with standard chemotherapy (Wedge et al. (Deja et al. 2014; Puchades-Carrasco et al. 2016). In the 2011) and to characterize KRAS mutants in NSCLC cells above-mentioned studies, changes in many distinct groups (Brunelli et al. 2014). The current study provided new data of metabolites were reported, such as amino acids, carbohy- on metabolites, which could translate to the improvement drates, organic acids, fatty acids, and nucleotides. It should of lung cancer diagnosis and treatment. The study involved be noted that the results obtained in the previously pub- serum analysis of patients with newly diagnosed NSCLC lished metabolomic research were not always in agreement prior to initiation of a therapy and a matched control group. with each other. One of the discrepancies exists regarding Only patients with early stage disease (stage IA–IIB) were glutamate which was found elevated in blood of lung can- included to the study, representing an opportunity to define cer patients by Fahrmann et al. (2015) and Puchdes-Car- biomarkers for early lung cancer detection. rasco et al. (2016), whereas Hori et al. (2011) reported a decreased level. Another discrepancy is in the lactic acid level. According to Hori et al. (2011) and Musharraf et al. Methodology (2015), higher level of lactic acid occurred in the plasma of patients with lung cancer, whereas Chen et al. (2014) Reagents and materials observed its reduction in sera from lung cancer group as compared with controls. The possible reasons of the above- LC/MS grade acetonitrile, 0.1% formic acid in water, meth- mentioned discrepancies can be associated with differences anol, and acetone were purchased from Fischer Scientific in analytical methodologies, sample handling, or clinical Co. (Pittsburgh, PA, USA). Isotopically labeled stand- characteristics of patients. Therefore, it can be concluded ards: creatine-D3, L-leucine-D10, L-tryptophan-2,3,3-D3, that despite efforts made by many research groups, there and caffeine-D3 were obtained from CDN Isotopes Inc. 1 3 J Cancer Res Clin Oncol (Pointe-Claire, Quebec, Canada). Mixture of amino acids Table 1 Characteristics of the study participants was purchased from Sigma-Aldrich (St. Louis, MO, USA). Parameter Non-small-cell lung Control group cancer Study population N Total 50 25 The project was approved by the ethics committee at the Male 28 (56.0%) 14 (56.0%) Poznan University of Medical Sciences (Decision No. Female 22 (44.0%) 11 (44.0%) 200/13). All participants provided a written informed con- Age, year sent for inclusion to the study. Fifty lung cancer patients Mean 65 64 were recruited in the Department of Thoracic Surgery, Range 53–86 50–78 Poznan University of Medical Sciences. NSCLC diagno- BMI, kg/m2 sis was conducted by the histopathological examination of Mean 26.5 25.8 tissue samples. Patients with other tumors were excluded Range 17.6–33.9 21.1–34.6 from the study.