Metabolomics studies can be divided into two classes: targeted and non-targeted. Targeted analysis is used for the identification and quantification of pre-defined metabolites and can be used for quantitative as well as qualitative analysis
[83][37]. Non-targeted analysis consists of analysing all accessible metabolites in a given sample and is the first choice for cancer biomarker discovery studies
[84][38].
3.2. Sample Collection and Preparation
The collection of the sample, its preparation, and storage are the second step in the metabolomics study plan. The most common samples for conducting clinical metabolomics research are blood and urine
[88][39]. It is important to design the research based on metabolomics to reduce the influence of certain constituents such as age, gender, state of fasting, diet, physical activity, exercise, and the day and time of sample collection. Before starting the actual research, it is important to conduct a pilot study of healthy individuals and report it as part of the research to validate the results’ reproducibility. The samples (particularly plasma, serum, and urine) must be kept in various aliquots soon after collection to avoid the production of compounds from the many freeze–thaw cycles used for different metabolomics studies
[89][40]. The factors used for processing the sample, such as pH buffering and extraction, should also be uniform and follow standard operating procedures (SOPs)
[81,82,90,91][35][36][41][42]. The samples that are non-invasive in nature, such as blood or urine, are the best for regular clinical analysis
[85][43].
3.3. Analytical Techniques
3.3.1. LC–MS
The MS technique has the ability to isolate the intricate mixture of compounds for their detection and quantification with elevated sensitivity and specificity, and can also demonstrate information regarding molecular structures
[99][44]. MS separation techniques are essential for reducing sample complexity and minimising ionisation suppression effects
[100][45]. A preceding separation stage, such as high-performance liquid chromatography (HPLC), or ultra-performance liquid chromatography (UPLC), and capillary electrophoresis (CE), is frequently required.
3.3.2. GC–MS
GC–MS is a technique that combines great separation efficacy with sensitive, selective, and versatile mass evaluation and is suitable for comprehensive analysis. It is a combination of MS and GS that is used for the detection and quantification of a wide range of chemical compounds, such as natural products, blood, and urine. GS–MS is used in many fields of study, such as detecting drugs, amino acid evaluation, doping control, and the detection of natural materials like food products
[103][46]. EI, or electron ionization, is used for combining MS with GC in almost all the metabolomics applications that are based on GC. The EI–MS method works well for chemical compounds that do not change when heated and that are volatile and are separated by chromatography at high temperatures
[104][47].
3.3.3. NMR
NMR spectroscopy is a universal metabolite detection method that allows for direct analysis of samples with little sample preparation and simultaneous measurement of numerous types of tiny metabolites
[106,107][48][49]. However, it has limitations, such as high equipment costs, high maintenance costs, and decreased sensitivity
[108,109][50][51]. Mass spectrometry is better than NMR in several ways, although NMR has its own advantages.
3.4. Data Acquisition and Processing
When the metabolomics data are produced, it is important to ensure that they are reproducible
[89,110][40][52]. Quality standardisation and quality control are considered for the optimisation of the reproducibility of results. Data analysis and bioinformatics are used to process the data, which are then subjected to statistical analysis. There are two classical approaches to the statistical analysis of multivariate data: unsupervised learning and supervised learning. A popular unsupervised learning method is principal component analysis (PCA). The second main approach is supervised learning, such as with artificial neural networks (ANN), partial least squares discriminate analysis (PLS-DA), etc., which can be used for excavating the data further to obtain the biomarkers
[111,112][53][54]. The discovery process of biomarkers can be driven through supervised models that can be linked with clinical results, histopathological scores, and various other omics data. It is important to test the supervised models with precise internal cross-validation processes or external tests to obtain trusted biomarkers and models and to decrease the chances of data overfitting
[113][55].
3.5. Metabolites Identification: Biomarker Discovery and Validation
Profiling the metabolites in each biological entity is incomplete without accurate data measurement and precise interpretation. To identify the features of potent spectral biomarkers, attempts are made to recognise the unidentified spectral biomarkers. The peaks can be identified with the help of public metabolomics databases and in-house spectral databases such as the Golm database, LIPID MAPS, human metabolome database (HMDB), METLIN database, etc. Following the identification of metabolomics biomarkers, additional experiments are required to validate or test the biomarkers
[82,112,114][36][54][56].
4. Applications of Metabolomics in B-NHL
4.1. Discovering Targeted Therapies Based on Metabolomics
Metabolism in B-NHL plays a crucial role in established therapeutic approaches. Antimetabolites were the name given to the chemical compounds that were first used to treat cancer. The reason for choosing this name was that these compounds were found to resemble endogenous metabolites in their chemical structure and disrupt the process of normal metabolism. In comparison to other omics, metabolomics is best for evaluating the potential of these cancer treatment regimens. The study was carried out to discover whether the therapies could cause alterations in the metabolic pathways and detect the pharmacokinetics of drugs simultaneously or not. In the coming time, it will become crucial to combine the study of pharmacometabolomics with other biological systems knowledge, such as mRNA, genetics, miRNA, and imaging. This will help in determining the correlation of the metabolomics response with the cancer stage, undesirable incidents, and the growth or recession of the tumour. The study of pharmacometabolomics is capable of monitoring a patient’s metabolic response to a drug; thus, it is very interesting to use metabolomics in detecting cancerous growth, prognosis, and therapy management
[82][36].
4.2. Determining B-NHL Diagnostic and Prognostic Biomarkers
A recent metabolomics study suggested a methodology for discovering novel biomarkers that can be used for the diagnosis and characterisation of various lymphoma subtypes. The GC–MS method was used for the investigation and evaluation of plasma samples taken from individuals with different subtypes of lymphomas. The results showed a significant prevalence of elaidic acid and hypoxanthine (HX) in patients suffering from Hodgkin’s lymphoma, MM, CLL, and DLBCL compared with healthy control individuals in all the study groups
[50][57].
4.3. Determining the Lymphomagenesis Risk Factors
Genetic mutations accumulate sequentially during tumour development, eventually resulting in malignant tumours. However, it has also been shown that metabolic processes and inflammatory factors indirectly contribute to the development of the tumour
[11][58]. In their study, Pettersena et al. proved that the cell line of B-cell lymphoma surrounds numerous amplified genomic uracil concentrations in comparison with non-lymphoma cell lines or normal lymphocytes. They utilised a method based on liquid chromatography combined with mass spectrometry (LC/MS) for quantifying the genomic sequence of 2-deoxyuridine and proving their study. In harmony with uracil generated by activation-induced cytidine deaminase (AID), they discovered a distinctive mutational signature of an AID hotspot in the lymphoma area where there was clustered mutation. They also presented an important revelation about the expression of SMUG1 and uracil-DNA glycosylases UNG along with the excision capacity of uracil by stating its negative correlation with the concentration of genomic uracil, which somewhat decreased the AID effect
[129][59]. Another study was also conducted on the metabolomic pattern of Burkitt lymphoma that was induced by MYC glucose deprivation, as well as hypoxic and aerobic conditions. They used a [U-13C, 15N]-glutamine tracer to detect glutamine import and metabolism via the TCA cycle under hypoxia conditions and discovered that glutamine is significantly precipitated to citrate carbons. The deficiency of glucose leads to the significant augmentation of citrate, fumarate, and glutamine-derived malate. Their arrangements showed a different pathway for the generation of energy called glutaminolysis, which is associated with the glucose-independent TCA cycle. Under the conditions of hypoxia and scarcity of glucose, the critical role of glutamine in the proliferation of cells makes them susceptible to BPTES (glutaminase inhibitors), which in turn can be used for treating tumours
[130][60].