In this study, ¹H NMR-based metabonomics has been applied, for the first time to our knowledge, to investigate lung cancer metabolic signatures in urine, aiming at assessing the diagnostic potential of this approach and gaining novel insights into lung cancer metabolism and systemic effects. Urine samples from lung cancer patients (n = 71) and a control healthy group (n = 54) were analyzed by high resolution ¹H NMR (500 MHz), and their spectral profiles subjected to multivariate statistics, namely, Principal Component Analysis (PCA), Partial Least Squares Discriminant Analysis (PLS-DA), and Orthogonal Projections to Latent Structures (OPLS)-DA. Very good discrimination between cancer and control groups was achieved by multivariate modeling of urinary profiles. By Monte Carlo Cross Validation, the classification model showed 93% sensitivity, 94% specificity and an overall classification rate of 93.5%. The possible confounding influence of other factors, namely, gender and age, have also been modeled and found to have much lower predictive power than the presence of the disease. Moreover, smoking habits were found not to have a dominating influence over class discrimination. The main metabolites contributing to this discrimination, as highlighted by multivariate analysis and confirmed by spectral integration, were hippurate and trigonelline (reduced in patients), and β-hydroxyisovalerate, α-hydroxyisobutyrate, N-acetylglutamine, and creatinine (elevated in patients relatively to controls). These results show the valuable potential of NMR-based metabonomics for finding putative biomarkers of lung cancer in urine, collected in a minimally invasive way, which may have important diagnostic impact, provided that these metabolites are found to be specifically disease-related.

This work aims at characterizing the metabolic profile of human lung cancer, to gain new insights into tumor metabolism and to identify possible biomarkers with potential diagnostic value in the future. Paired samples of tumor and noninvolved adjacent tissues from 12 lung tumors have been directly analyzed by (1)H HRMAS NMR (500/600 MHz) enabling, for the first time to our knowledge, the identification of over 50 compounds. The effect of temperature on tissue stability during acquisition time has also been investigated, demonstrating that analysis should be performed within less than two hours at low temperature (277 K), to minimize glycerophosphocholine (GPC) and phosphocholine (PC) conversion to choline and reduce variations in some amino acids. The application of Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) to the standard 1D (1)H spectra resulted in good separation between tumor and control samples, showing that inherently different metabolic signatures characterize the two tissue types. On the basis of spectral integration measurements, lactate, PC, and GPC were found to be elevated in tumors, while glucose, myo-inositol, inosine/adenosine, and acetate were reduced. These results show the valuable potential of HRMAS NMR-metabonomics for investigating the metabolic phenotype of lung cancer.

Lung cancer is one of the most prevalent and fatal types of cancer, with average 5-year survival rates lower than 15%, 1,2 which is mainly due to the advanced stage at which lung tumors are usually diagnosed. Indeed, when lung cancer is detected before metastasizing to lymph nodes or distant sites, the 5-year survival rates increase drastically to 60À80%, thus stressing the importance of early diagnosis. However, the majority of patients show no signs or symptoms during the initial phases of neoplastic growth, hindering early detection and the possibility of curative surgical treatment. Moreover, radiological tests, such as computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET), which would allow the detection of initial cancer lesions, are not suitable for general screening of the population, mainly due to their high costs. Therefore, new methods that can aid in the early detection of lung cancer and contribute to improved prognosis are greatly needed.The search for metabolic markers of cancer in human tissues and biofluids has been the focus of a number of metabonomic studies in recent years. 3,4 In particular, metabolic profiling of blood plasma or serum has been increasingly used to unveil metabolic alterations associated with different cancer types, such as breast, 5À7 kidney, 8À10 liver, 11À13 prostate, 14À16 colorectal, 17À20 oral, 21,22 pancreatic, 23,24 esophageal, 25 and bone 26 cancers. In the case of lung cancer, only a few studies focusing on plasma or serum metabolic profiling have been recently reported. Maeda and co-workers proposed that the differences in the plasma amino acid profiles between healthy controls and non-small-cell lung cancer (NSCLC) patients, as assessed by targeted liquid chromatography coupled to mass spectrometry (LCÀMS), could potentially be useful for screening NSCLC. 27 In another MS study of specific compounds, namely, lysophosphatidylcholines (lysoPC), abnormal levels of lysoPC isomers with different fatty acyl positions were found in the plasma of lung cancer patients compared to controls. 28 Using a more global profiling approach, Jordan and colleagues reported the NMR analysis of paired tissue and serum samples from 14 subjects with two different lung cancer histological types (adenocarcinoma and squamous cell

Lung tumour subtyping, particularly the distinction between adenocarcinoma (AdC) and squamous cell carcinoma (SqCC), is a critical diagnostic requirement. In this work, the metabolic signatures of lung carcinomas were investigated through (1)H NMR metabolomics, with a view to provide additional criteria for improved diagnosis and treatment planning. High Resolution Magic Angle Spinning Nuclear Magnetic Resonance (NMR) spectroscopy was used to analyse matched tumour and adjacent control tissues from 56 patients undergoing surgical excision of primary lung carcinomas. Multivariate modeling allowed tumour and control tissues to be discriminated with high accuracy (97% classification rate), mainly due to significant differences in the levels of 13 metabolites. Notably, the magnitude of those differences were clearly distinct for AdC and SqCC: major alterations in AdC were related to phospholipid metabolism (increased phosphocholine, glycerophosphocholine and phosphoethanolamine, together with decreased acetate) and protein catabolism (increased peptide moieties), whereas SqCC had stronger glycolytic and glutaminolytic profiles (negatively correlated variations in glucose and lactate and positively correlated increases in glutamate and alanine). Other tumour metabolic features were increased creatine, glutathione, taurine and uridine nucleotides, the first two being especially prominent in SqCC and the latter in AdC. Furthermore, multivariate analysis of AdC and SqCC profiles allowed their discrimination with a 94% classification rate, thus showing great potential for aiding lung tumours subtyping. Overall, this study has provided new, clear evidence of distinct metabolic signatures for lung AdC and SqCC, which can potentially impact on diagnosis and provide important leads for future research on novel therapeutic targets or imaging tracers.