Emily G. Armitage

Metabolomics in cancer biomarker discovery: current trends and future perspectives.

Cancer is one of the most devastating human diseases that causes a vast number of mortalities worldwide each year. Cancer research is one of the largest fields in the life sciences and despite many astounding breakthroughs and contributions over the past few decades, there is still a considerable amount to unveil on the function of cancer. It is well known that cancer metabolism differs from that of normal tissue and an important hypothesis published in the 1950s by Otto Warburg proposed that cancer cells rely on anaerobic metabolism as the source for energy, even under physiological oxygen levels. Following this, cancer central carbon metabolism has been researched extensively and beyond respiration, cancer has been found to involve a wide range of metabolic processes, and many more are still to be unveiled. Studying cancer through metabolomics could reveal new biomarkers for cancer that could be useful for its future prognosis, diagnosis and therapy. Metabolomics is becoming an increasingly popular tool in the life sciences since it is a relatively fast and accurate technique that can be applied with either a particular focus or in a global manner to reveal new knowledge about biological systems. There have been many examples of its application to reveal potential biomarkers in different cancers that have employed a range of different analytical platforms. In this review, approaches in metabolomics that have been employed in cancer biomarker discovery are discussed and some of the most noteworthy research in the field is highlighted.

Controlling the quality of metabolomics data: new strategies to get the best out of the QC sample

The type and use of quality control (QC) samples is a ‘hot topic’ in metabolomics. QCs are not novel in analytical chemistry; however since the evolution of using QCs to control the quality of data in large scale metabolomics studies (first described in 2011), the need for detailed knowledge of how to use QCs and the effects they can have on data treatment is growing. A controlled experiment has been designed to illustrate the most advantageous uses of QCs in metabolomics experiments. For this, samples were formed from a pool of plasma whereby different metabolites were spiked into two groups in order to simulate biological biomarkers. Three different QCs were compared: QCs pooled from all samples, QCs pooled from each experimental group of samples separately and QCs provided by an external source (QC surrogate). On the experimentation of different data treatment strategies, it was revealed that QCs collected separately for groups offers the closest matrix to the samples and improves the statistical outcome, especially for biomarkers unique to one group. A novel quality assurance plus procedure has also been proposed that builds on previously published methods and has the ability to improve statistical results for QC pool. For this dataset, the best option to work with QC surrogate was to filter data based only on group presence. Finally, a novel use of recursive analysis is portrayed that allows the improvement of statistical analyses with respect to the ratio between true and false positives.

Influence of Missing Values Substitutes on Multivariate Analysis of Metabolomics Data

Missing values are known to be problematic for the analysis of gas chromatography-mass spectrometry (GC-MS) metabolomics data. Typically these values cover about 10%–20% of all data and can originate from various backgrounds, including analytical, computational, as well as biological. Currently, the most well known substitute for missing values is a mean imputation. In fact, some researchers consider this aspect of data analysis in their metabolomics pipeline as so routine that they do not even mention using this replacement approach. However, this may have a significant influence on the data analysis output(s) and might be highly sensitive to the distribution of samples between different classes. Therefore, in this study we have analysed different substitutes of missing values namely: zero, mean, median, k-nearest neighbours (kNN) and random forest (RF) imputation, in terms of their influence on unsupervised and supervised learning and, thus, their impact on the final output(s) in terms of biological interpretation. These comparisons have been demonstrated both visually and computationally (classification rate) to support our findings. The results show that the selection of the replacement methods to impute missing values may have a considerable effect on the classification accuracy, if performed incorrectly this may negatively influence the biomarkers selected for an early disease diagnosis or identification of cancer related metabolites. In the case of GC-MS metabolomics data studied here our findings recommend that RF should be favored as an imputation of missing value over the other tested methods. This approach displayed excellent results in terms of classification rate for both supervised methods namely: principal components-linear discriminant analysis (PC-LDA) (98.02%) and partial least squares-discriminant analysis (PLS-DA) (97.96%) outperforming other imputation methods.

Searching for urine biomarkers of bladder cancer recurrence using a liquid chromatography–mass spectrometry and capillary electrophoresis–mass spectrometry metabolomics approach

The incidence and rate of recurrence of bladder cancer is high, particularly in developed countries, however current methods for diagnosis are limited to detecting high-grade tumours using often invasive methods. A panel of biomarkers to characterise tumours of different grades that could also distinguish between patients exhibiting the disease with first incidence or recurrence could be useful for bladder cancer diagnostics. In this study, potential metabolic biomarkers have been discovered through mass spectrometry based metabolomics of urine. Pre-treatment urine samples were collected from 48 patients diagnosed of urothelial bladder cancer. Patients were followed-up through the hospital pathological charts to identify whether and when the disease recurred or progressed. Subsequently, they were classified according to whether or not they suffered a tumour recurrence (recurrent or stable) as well as their risk group according to tumour grade and stage. Identified metabolites have been analysed in terms of disease characteristics (tumour stage and recurrence) and have provided an insight into bladder cancer progression. Using both liquid chromatography and capillary electrophoresis-mass spectrometry, a total of 27 metabolite features were highlighted as significantly different between patient groups. Some, for example histidine, phenylalanine, tyrosine and tryptophan have been previously linked with bladder cancer, however until now their connection with bladder cancer progression has not been previously reported. The candidate biomarkers revealed in this study could be useful in the clinic for diagnosis of bladder cancer and, through characterising the stage of the disease, could also be useful in prognostics.

A novel untargeted metabolomics correlation-based network analysis incorporating human metabolic reconstructions

BackgroundMetabolomics has become increasingly popular in the study of disease phenotypes and molecular pathophysiology. One branch of metabolomics that encompasses the high-throughput screening of cellular metabolism is metabolic profiling. In the present study, the metabolic profiles of different tumour cells from colorectal carcinoma and breast adenocarcinoma were exposed to hypoxic and normoxic conditions and these have been compared to reveal the potential metabolic effects of hypoxia on the biochemistry of the tumour cells; this may contribute to their survival in oxygen compromised environments. In an attempt to analyse the complex interactions between metabolites beyond routine univariate and multivariate data analysis methods, correlation analysis has been integrated with a human metabolic reconstruction to reveal connections between pathways that are associated with normoxic or hypoxic oxygen environments.ResultsCorrelation analysis has revealed statistically significant connections between metabolites, where differences in correlations between cells exposed to different oxygen levels have been highlighted as markers of hypoxic metabolism in cancer. Network mapping onto reconstructed human metabolic models is a novel addition to correlation analysis. Correlated metabolites have been mapped onto the Edinburgh human metabolic network (EHMN) with the aim of interlinking metabolites found to be regulated in a similar fashion in response to oxygen. This revealed novel pathways within the metabolic network that may be key to tumour cell survival at low oxygen. Results show that the metabolic responses to lowering oxygen availability can be conserved or specific to a particular cell line. Network-based correlation analysis identified conserved metabolites including malate, pyruvate, 2-oxoglutarate, glutamate and fructose-6-phosphate. In this way, this method has revealed metabolites not previously linked, or less well recognised, with respect to hypoxia before. Lactate fermentation is one of the key themes discussed in the field of hypoxia; however, malate, pyruvate, 2-oxoglutarate, glutamate and fructose-6-phosphate, which are connected by a single pathway, may provide a more significant marker of hypoxia in cancer.ConclusionsMetabolic networks generated for each cell line were compared to identify conserved metabolite pathway responses to low oxygen environments. Furthermore, we believe this methodology will have general application within metabolomics.

Missing value imputation strategies for metabolomics data.

The origin of missing values can be caused by different reasons and depending on these origins missing values should be considered differently and dealt with in different ways. In this research, four methods of imputation have been compared with respect to revealing their effects on the normality and variance of data, on statistical significance and on the approximation of a suitable threshold to accept missing data as truly missing. Additionally, the effects of different strategies for controlling familywise error rate or false discovery and how they work with the different strategies for missing value imputation have been evaluated. Missing values were found to affect normality and variance of data and k‐means nearest neighbour imputation was the best method tested for restoring this. Bonferroni correction was the best method for maximizing true positives and minimizing false positives and it was observed that as low as 40% missing data could be truly missing. The range between 40 and 70% missing values was defined as a “gray area” and therefore a strategy has been proposed that provides a balance between the optimal imputation strategy that was k‐means nearest neighbor and the best approximation of positioning real zeros.

Metabolomics as a tool for drug discovery and personalised medicine. A review

Studying the effects of drugs on the metabolome constitutes a huge part of the metabolomics discipline. Whether the approach is associated with drug discovery (altered pathways due to the disease that provide future targets and information into the mechanism of action or resistance, etc.) or pharmacometabolomics (studying the outcome of treatment), there have been many aspiring published articles in this area. With specific experimental design, including fingerprinting analysis with different analytical platforms in a non-targeted way, the approach is advancing towards the discovery of markers for the implication of personalised medicine, while also providing information that could help to improve the efficacy and reduce the side effects associated with a treatment. In this review, the evolution of pharmacometabolomics from other areas of drug efficacy metabolomics studies is explored.

Looking into aqueous humor through metabolomics spectacles - exploring its metabolic characteristics in relation to myopia.

Aqueous humor is the transparent fluid found in the anterior chamber of the eye that provides the metabolic requirements to the avascular tissues surrounding it. Despite the fact that metabolomics could be a powerful tool in the characterization of this biofluid and in revealing metabolic signatures of common ocular diseases such as myopia, it has never to our knowledge previously been applied in humans. In this research a novel method for the analysis of aqueous humor is presented to show its application in the characterization of this biofluid using CE-MS. The method was extended to a dual platform method (CE-MS and LC-MS) in order to compare samples from patients with different severities of myopia in order to explore the disease from the metabolic phenotype point of view. With this method, a profound knowledge of the metabolites present in human aqueous humor has been obtained: over 40 metabolites were reproducibly and simultaneously identified from a low volume of sample by CE-MS, including among others, a vast number of amino acids and derivatives. When this method was extended to study groups of patients with high or low myopia in both CE-MS and LC-MS, it has been possible to identify over 20 significantly different metabolite and lipid signatures that distinguish patients based on the severity of myopia. Among these, the most notable higher abundant metabolites in high myopia were aminooctanoic acid, arginine, citrulline and sphinganine while features of low myopia were aminoundecanoic acid, dihydro-retinoic acid and cysteinylglycine disulfide. This dual platform approach offered complementarity such that different metabolites were detected in each technique. Together the experiments presented provide a whelm of valuable information about human aqueous humor and myopia, proving the utility of non-targeted metabolomics for the first time in analyzing this type of sample and the metabolic phenotype of this disease.

A Single In-Vial Dual Extraction Strategy for the Simultaneous Lipidomics and Proteomics Analysis of HDL and LDL Fractions

A single in-vial dual extraction (IVDE) procedure for the subsequent analysis of lipids and proteins in the high-density lipoprotein (HDL) and low-density lipoprotein (LDL) fractions derived from the same biological sample is presented. On the basis of methyl-tert-butyl ether (MTBE) extraction, IVDE leads to the formation of three phases: a protein pellet at the bottom, an aqueous phase with polar compounds, and an ether phase with lipophilic compounds. After sample extraction, performed within a high-performance liquid chromatography vial insert, the ether phase was directly injected for lipid fingerprinting, while the protein pellet, after evaporation of the remaining sample, was used for proteomics analysis. Human HDL and LDL isolates were used to test the suitability of the IVDE methodology for lipid and protein analysis from a single sample in terms of data quality and matching composition to that of HDL and LDL. Subsequently, HDL and LDL fractions isolated from ApoE-KO and wild-type mice were used to validate the capacity of IVDE for revealing changes in lipid and protein abundance. Results indicate that IVDE can be successfully used for the subsequent analysis of lipids and proteins with the advantages of time saving, simplicity, and reduced sample amount.

Metabolic profiling reveals potential metabolic markers associated with Hypoxia Inducible Factor-mediated signalling in hypoxic cancer cells

Hypoxia inducible factors (HIFs) plays an important role in oxygen compromised environments and therefore in tumour survival. In this research, metabolomics has been applied to study HIFs metabolic function in two cell models: mouse hepatocellular carcinoma and human colon carcinoma, whereby the metabolism has been profiled for a range of oxygen potentials. Wild type cells have been compared to cells deficient in HIF signalling to reveal its effect on cellular metabolism under normal oxygen conditions as well as low oxygen, hypoxic and anoxic environments. Characteristic responses to hypoxia that were conserved across both cell models involved the anti-correlation between 2-hydroxyglutarate, 2-oxoglutarate, fructose, hexadecanoic acid, hypotaurine, pyruvate and octadecenoic acid with 4-hydroxyproline, aspartate, cysteine, glutamine, lysine, malate and pyroglutamate. Further to this, network-based correlation analysis revealed HIF specific pathway responses to each oxygen condition that were also conserved between cell models. From this, 4-hydroxyproline was revealed as a regulating hub in low oxygen survival of WT cells while fructose appeared to be in HIF deficient cells. Pathways surrounding these hubs were built from the direct connections of correlated metabolites that look beyond traditional pathways in order to understand the mechanism of HIF response to low oxygen environments.