Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Metabolic profiling of human melanoma cell lines with high and low metastatic capacity by 1H-NMR spectroscopy

  • Nima Rezvani Kakhki,

    Roles Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Visualization, Writing – original draft

    Affiliations Department of Pharmaceutical Chemistry, Semmelweis University, Budapest, Hungary, Center for Pharmacology and Drug Research & Development, Semmelweis University, Budapest, Hungary

    ⨯
  • Zita Hegedűs,

    Roles Investigation, Writing – original draft

    Affiliation National Korányi Institute of Pulmonology, Budapest, Hungary

    ⨯
  • József Tóvári,

    Roles Conceptualization, Funding acquisition, Methodology, Resources, Supervision, Writing – review & editing

    Affiliation Department of Experimental Pharmacology and the National Tumor Biology Laboratory, National Institute of Oncology, Budapest, Hungary

    ⨯
  • Arash Mirzahosseini,

    Roles Data curation, Formal analysis, Funding acquisition

    Affiliations Department of Pharmaceutical Chemistry, Semmelweis University, Budapest, Hungary, Center for Pharmacology and Drug Research & Development, Semmelweis University, Budapest, Hungary

    ⨯
  • Béla Noszál,

    Roles Conceptualization, Resources, Supervision, Writing – review & editing

    Affiliations Department of Pharmaceutical Chemistry, Semmelweis University, Budapest, Hungary, Center for Pharmacology and Drug Research & Development, Semmelweis University, Budapest, Hungary

    ⨯
  • Márta Kraszni

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Supervision, Writing – review & editing

    kraszni.marta@semmelweis.hu

    Affiliations Department of Pharmaceutical Chemistry, Semmelweis University, Budapest, Hungary, Center for Pharmacology and Drug Research & Development, Semmelweis University, Budapest, Hungary

    ⨯

Abstract

Background

Melanoma is one of the most aggressive forms of skin cancer due to its high metastatic potential and mortality rate. Although understanding of metabolic reprogramming in melanoma has advanced, the connection between metabolic alterations and metastatic capacity remains incomplete.

Aim

This study aimed to characterize the metabolic profiles of human melanoma cell lines with high (HT168-M1) and low (WM983B) metastatic potential, and to compare them with each other and also with the metabolic profile of normal human fibroblasts (MRC-5), in order to identify key metabolites and metabolic pathways associated with metastatic behavior.

Methods

Non-targeted metabolomic profiling using ¹H-NMR spectroscopy was applied to hydrophilic extracts of the three cell lines. Multivariate statistical analyses (PCA and PLS-DA) were used to identify discriminating metabolites, and pathway analysis was performed to determine altered metabolic networks.

Results

Several metabolic pathways were significantly altered in melanoma cells compared to fibroblasts, including starch and sucrose metabolism, alanine, aspartate and glutamate metabolism, and glutathione metabolism. Metabolites showing more than two-fold differences included elevated UDP-glucose, ATP, glycerophosphocholine, GTP, creatine and glutathione in the melanoma cells, and reduced glucose, glutamine and 1-methylnicotinamide in fibroblasts. Comparison of the metabolites of melanoma cell lines with differing metastatic potential revealed changes in taurine and hypotaurine, β-alanine-, glutathione-, and amino acid metabolism. Metabolites showing the largest concentration changes were UDP-glucose, glutathione, NAD+, alanine and β-alanine.

Conclusion

Metabolomic profiling revealed distinct metabolic reprogramming between melanoma and normal fibroblasts, characterized by enhanced glycolysis and glutathione-dependent antioxidant defense. Highly metastatic melanoma cells demonstrated stronger redox adaptation and altered amino acid utilization, with elevated glutathione and glutamate and reduced NAD⁺ and pyruvate, indicating a metabolic shift toward oxidative stress resistance.

Citation: Kakhki NR, Hegedűs Z, Tóvári J, Mirzahosseini A, Noszál B, Kraszni M (2026) Metabolic profiling of human melanoma cell lines with high and low metastatic capacity by 1H-NMR spectroscopy. PLoS One 21(7): e0352639. https://doi.org/10.1371/journal.pone.0352639

Editor: Roy P. Planalp, University of New Hampshire, UNITED STATES OF AMERICA

Received: December 12, 2025; Accepted: June 11, 2026; Published: July 1, 2026

Copyright: © 2026 Kakhki et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: This research was financially supported by the National Tumor Biology Laboratory under the National Laboratories Program (2022-2.1.1-NL-2022-00010), the Hungarian Thematic Excellence Program (TKP2021-EGA-44 to J.T. and TKP2021-EGA-24 to A.M.), and the National Research, Development and Innovation Office’s Hungarian Scientific Research Fund (NKFIH-OTKA) grants K147410 (J.T.). The work was also funded by the New National Excellence Program of the Ministry for Culture and Innovation from the source of the National Research, Development and Innovation Fund (ÚNKP-23-5-SE-1 to A.M.) (https://nkfih.gov.hu). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Melanoma, also known as malignant melanoma or cutaneous melanoma, is a type of cancer that originates in melanocytes [1]. It is estimated that there will be over 110,000 new melanoma cases in the US in 2026, accounting for around 5.3% of all cancer cases. Epidemiological data indicate that the 5-year relative survival rate for non-metastatic, localised melanoma is over 99%. However, this drops to 76% for regional metastatic cases and 35% for distant metastatic cases [2]. Despite ongoing, intense research into metastasis, anti-metastatic therapies are either lacking or show limited success [3]. Therefore, a deeper understanding of how metastatic cancer cells work is still required.

Tumor cells generate energy primarily through glycolysis, rather than oxidative phosphorylation and the tricarboxylic acid (TCA) cycle, even when there is sufficient oxygen present (Warburg effect) [4]. Although glycolysis produces smaller amount of ATP than oxidative phosphorylation (OXPHOS), it provides energy more quickly to meet the high energy demands of rapidly proliferating cells [5]. Metastasizing cancer cells dynamically adapt their metabolism to the changing environment during the metastatic cascade, which is crucial for their survival [6]. In metastatic cancer, the transcription factors HIF-1α (hypoxia-inducible factor 1 alpha) and c-Myc play an important role in the upregulation of many enzymes involved in glucose metabolism. These include serum lactate dehydrogenase (LDH), which catalyzes the conversion of pyruvate to lactate, resulting in high lactate levels in cancer cells. Elevated LDH is associated with a poor prognosis in melanoma and several other cancers [7–10]. However, further progression requires increased mitochondrial function, as well as increased glutaminolysis and fatty acid oxidation [11]. In addition to being an energy source, glutamine is a precursor for the synthesis of glutathione, which is essential for maintaining redox homeostasis. It also serves as a source of carbon and nitrogen for the biosynthesis of amino acids, nucleotides, and lipids [12,13]. The metabolism of several other amino acids is also disturbed in metastatic cells. Kim et al. observed a progressive increase in several amino acids, including glycine, pyroglutamic acid, cysteine, ornithine, tyrosine and lysine, as the metastatic potential of melanoma cell lines increased [14].

As changes in metabolite levels are a molecular consequence of changes in cellular activity, metabolic profiling is an important tool for discovering biomarkers that can be used to diagnose tumor progression [15–18]. Of the analytical techniques used to monitor metabolic changes in different types of cancer, HPLC-MS, GC-MS and NMR spectroscopy have been employed [19–21]. As each platform has its own advantages and disadvantages, they are considered complementary techniques in metabolomics research. LC–MS/MS-based analytical approaches are an integral component of contemporary clinical and translational oncology, particularly with regard to metabolite quantification and pharmacological monitoring [22,23]. LC-MS is highly sensitive, making it suitable for analysing low-concentration metabolites. In contrast, high-field and high-resolution NMR spectroscopy offers lower sensitivity but higher reproducibility. Moreover, NMR-based metabolomics has several additional advantages, including its non-destructive nature, minimal sample preparation requirements, and inherently quantitative signal response. Furthermore, NMR offers robust structural information, facilitating reliable metabolite identification. These features make NMR particularly suitable for comparative metabolic profiling, despite the higher sensitivity typically achieved by MS-based approaches [24]. For the evaluation of NMR data, different multivariate statistical methods are complementarily needed. Unsupervised principal component analysis (PCA) is commonly used as an exploratory method to detect potential outliers and visualize clustering patterns in datasets. Partial least squares (PLS) regression, particularly when combined with discriminant analysis (PLS-DA), identifies and models the metabolite patterns that most effectively distinguish between predefined groups, focusing on variation directly related to class differences. The importance of the variables can be assessed using variable importance in projection (VIP) scores, which show how much each variable contributes to the model [25].

Despite increasing knowledge of metabolic reprogramming in melanoma, the relationship between metabolic alterations and metastatic potential remains incompletely understood. While several studies have identified metabolic differences between cancerous and non-cancerous cells, fewer have systematically compared melanoma cell lines with distinct metastatic capacities under controlled conditions [14,26–30]. Moreover, the metabolic characterization of specific melanoma models such as HT168-M1 and WM983B remains limited, particularly using high-resolution 1H-NMR-based metabolomics.

Therefore, the aim of the present study was to characterize the metabolic profiles of melanoma cell lines with high (HT168-M1) and low (WM983B) metastatic potential, and to compare them with normal human fibroblasts (MRC-5) as a physiological reference. Fibroblasts are increasingly recognized as metabolically active and highly responsive cells that regulate tissue homeostasis through interconnected signaling and mechanobiological pathways, and their established role in modulating tumor metabolism and the microenvironment further supports their suitability as physiologically relevant controls in cancer metabolomics studies [31,32]. The WM983B and HT168-M1 cell lines are both derived from lymph node metastases, but the former has much lower metastatic potential. The HT168-M1 human melanoma cell line is the derivative of the A2058 cell line and has an extremely high liver-colonizing capacity [33–36]. Both cell lines express the BRAF V600E mutation and thus exhibit increased HIF-1α expression.

By applying 1H NMR-based metabolomics and multivariate statistical analysis, we sought to identify key metabolites and pathways associated with metastatic behavior, thereby contributing to a better understanding of melanoma progression and potential metabolic vulnerabilities.

Materials and methods

Cell cultures and cell collections

The HT168-M1 human melanoma line (RRID: CVCL_2H39) is derivatives from the A2058 cell line [35] with in vivo selection for high liver metastatic capacity. WM983B (RRID: CVCL_6809) melanoma cell lines were gifts from M. Herlyn (Wistar Institute, Philadelphia, PA). MRC-5 (RRID: CCL-171) was obtained from ATCC. The human melanoma cells were cultured in RPMI-1640 culture media (Sigma-Aldrich Ltd., Budapest, Hungary), and the MRC-5 human fibroblast cells were grown in DMEM (Dulbecco’s modified Eagle’s medium, Sigma-Aldrich Ltd., Budapest, Hungary) culture media with 10% FBS (Fetal bovine serum) and 1% Penicillin/Streptomycin. The culture medium was removed from the culture flasks (4 pieces of confluent T-75 flask per cell line), then cells were washed up quickly twice with ice-cold PBS (phosphate buffered saline). The residual PBS was removed by vacuum. Then, 3 ml of ice-cold methanol was added to the cells, and cell-scraper was used for cell collection. The methanol, containing the cells was pipetted into a 15 ml centrifuge tube. To ensure statistical reliability, biological replicates were prepared from each cell line. A total of 43 cell cultures were produced, comprising 16 HT168-M1, 16 WM983B, and 11 MRC5 cultures. All cells were cultured in the Department of Experimental Pharmacology, National Institute of Oncology, Budapest, Hungary.

Extraction of metabolites

The quenched cells were suspended in methanol, frozen in liquid nitrogen, and then thawed using a sonicator water bath (Bransonic Ultrasonic cleaner, Branson 1510-MT, Branson Ultrasonic corp., Danbury, CT, US) at room temperature to rupture the cell membrane and release metabolites from the cytosol. The freeze-thaw cycle was repeated three times. The polar metabolites were extracted from the cultured cells using a mixture of methanol, chloroform and water in a ratio of 10:10:9, with a volume of 15 ml. The same sonicator was used as before to mix the solvents. In addition, the mixtures were vortexed for 30 seconds using a classic vortex mixer (Velp Scientifica Srl, Usmate, Italy). The phases were then separated by centrifugation at 10,000 x g for 10 minutes using Sigma 2-16P benchtop centrifuge (Sigma Laborzentrifugen GmbH, Germany). After centrifugation, the aqueous/methanol phase containing the polar metabolites was transferred to 15 ml centrifuge tubes, and most of the solvent was evaporated under a stream of nitrogen. The remaining solvent was then removed by freeze drying (Scanvac CoolSafe 55−4 LaboGene ApS Lynge, Denmark).

NMR measurements

The extracted and dried material was dissolved in 800 μL of D2O (99.96% D, VWR International Ltd., Budapest, Hungary), and 20 μL of a tenfold diluted TSP solution (0.75 wt.% trimethylsilyl propionic acid sodium salt, 99.9% D, Sigma-Aldrich Ltd., Budapest, Hungary) was added as a chemical shift reference standard. To prevent bacterial growth, 0.2 mg of sodium azide (Sigma-Aldrich Ltd., Budapest, Hungary) was used. The pH of the final solutions was adjusted to 7.00 ± 0.02 using diluted DCl and NaOD solutions, with the pH being measured using an Orion Star A211 benchtop pH meter (Thermo Fisher Scientific Inc., Waltham, MA, US) equipped with a Metrohm 6.0204.100 combined pH glass electrode, which was calibrated by NBS (National Bureau of Standards) buffer solutions. The solution was then centrifuged at 10,000 x g for 10 minutes using an Eppendorf MiniSpin microcentrifuge (Eppendorf AG, Hamburg, Germany). Then, 650 μL of the supernatant was transferred to a 5 mm NMR tube. NMR spectra were recorded at 25 °C using a Varian VNMRS spectrometer (Varian Inc., Palo Alto, USA) operating at a 1H frequency of 599.9 MHz. In the 1H-NMR experiments, presaturation of the residual water signal was applied. The spectra were obtained using the following acquisition parameters: a 90° flip angle, a relaxation time of 4.7 s, a spectral window of 7184 Hz, 32 K data points and 1024 scans.

NMR spectra processing and data transformation

NMR spectra were processed using the ACD NMR Processor Academic Edition 12.01 software (Advanced Chemistry Development Inc., Toronto, ON, Canada). The FIDs were digitized to 64k data points, and an exponential multiplication with a line broadening factor of 0.5 Hz was used. Following manual phase correction, automatic baseline correction (3rd order polynomial) was applied, and the spectra were referenced to TSP (0.00 ppm). Integration of the spectra was performed in 0.02 ppm bins from 0.8 ppm to 9.4 ppm excluding the region of the water signal (4.7–4.9 ppm).

1H-1H 2D TOCSY spectra were also recorded for the melanoma cell line extracts, using 100 ms mixing time, 80 transients and 256 increments. The spectra were digitized to 2048x2048 data points.

Statistical analysis

The statistical analysis of the NMR data was performed using Metaboanalyst 6.0 [37].

The binned data from the ACD NMR Processor program were extracted as a csv (comma separated value) file. A matrix containing data of 11–16 biological replicates of three different cell types was then built and imported into Metaboanalyst. First, 10% of the data was filtered out using the interquantile region variance filter. Then, the data were normalized using the sum of integrals, and autoscaling was applied to obtain a Gaussian distribution of the data [38]. Multivariate data analyses, including PCA and PLS-DA were performed on the data series, pairwise. Outliers were detected using PCA score plots of the first two PCs. The PLS-DA score plots of the first two components were then created to visualize the discrimination between the cell lines. The PLS-DA score plots were also generated, incorporating the binned data from all three cell lines. Finally, the VIP score plots of the first fifteen most important features were generated. Metabolites with VIP scores greater than 1.5 were considered important for discrimination purposes. Boxplots of the metabolite concentrations were used to visualize their relative concentrations in the cell lines, and fold change values were calculated to improve the interpretability of the results. The R2 and Q2 values obtained from a ten-fold cross validation process involving five components were used to evaluate the predictive ability of the PLS-DA models.

To aid the interpretation of the statistical results, the metabolite sets obtained in the PLS-DA analysis that showed discrimination, were subjected to Metabolic Pathway Analysis (MetPA) in MetaboAnalyst 6.0, using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway library for Homo sapiens [39].

Results

Identification of the metabolites

Fig 1 shows an example of a 1D 1H NMR spectrum of the extracted hydrophilic metabolites for each cell line. The spectral regions were assigned to the corresponding metabolites using the Human Metabolome Database (www.hmdb.ca) [40], as well as the recorded 2D 1H-1H TOCSY spectra of the melanoma cell extracts (S1–S3 Figs in S1 File). A total of 47 metabolites were identified; a list of these, along with their spectroscopic data, can be seen in Table 1.

thumbnail
Download:
Fig 1. Overlayed 1H NMR spectra of the studied cell lines.

(A) Aliphatic region. (B) Aromatic region (magnified twice compared to the aliphatic region). Y-axis: 1 – MRC5; 2 – WM983B; 3 – HT168-M1.

https://doi.org/10.1371/journal.pone.0352639.g001

thumbnail
Download:
Table 1. The 1H-NMR chemical shift and peak multiplicities of the metabolites identified in the NMR spectra of the studied cell lines.

https://doi.org/10.1371/journal.pone.0352639.t001

Multivariate data analysis

Statistical analysis was performed on a total of 43 cell line extracts (16 HT168-M1, 16 WM983B, 11 MRC5) using Metaboanalyst 6.0 [37]. After normalization by sum and scaling of the binned NMR data, PCA was first applied to detect possible outliers. Visual inspection of the PCA score plots, which were constructed using a pairwise analysis of the cell lines under study, revealed no outliers among the samples (S4 Fig in S1 File). Consequently, all 43 samples were included in the subsequent analysis.

Supervised PLS-DA analysis was used to classify the cell types in pairs. The PLS-DA scores plots clearly distinguish between the investigated cell line pairs (Fig 2A-C). PLS-DA scores plot was also created incorporating the data of all three cell lines. As can be seen in S5 Fig in S1 File, the melanoma cells are clearly distinct from one another. However, there is some overlap between HT168-M1 and MRC5, and greater overlap between WM983B and MRC5.

thumbnail
Download:
Fig 2. PLS-DA scores plots of the cell lines studied using pairwise analysis and VIP scores of the metabolites.

(A) Scores plots of MRC5 and WM983B cell lines. (B) Scores plots of MRC5 and HT168-M1 cell lines. (C) Scores plots of HT168-M1 and WM983B cell lines. (D) VIP scores of the most important metabolites distinguishing between MRC5 and WM983B cell lines. (E) VIP scores of the most important metabolites distinguishing between MRC5 and HT168-M1 cell lines. (F) VIP scores of the most important metabolites distinguishing between HT168-M1 and WM983B cell lines.

https://doi.org/10.1371/journal.pone.0352639.g002

The VIP score plots were also generated to show the important features (Fig 2D-F), revealing the first fifteen features that distinctively characterize cell types. As most metabolites have multiple signals in the 1H NMR spectrum, the metabolites assigned to the 0.02 bin spectral ranges (features) often appear repeatedly. Therefore, features with VIP scores greater than 1.5 were considered significant and are listed in S1-S3 Tables in S1 file. Overlapping metabolite signals mean that a single bin may be assigned to more than one metabolite. However, the multiple appearance of metabolites and selective integration areas (S4 Table in S1 file) make it possible to identify the metabolites responsible for spectral differences. These metabolites are listed in Tables 2–4, along with their fold change in concentration. The box plots of concentrations for the most important metabolites are shown in Figs 3–5.

thumbnail
Download:
Table 2. Fold change of concentrations (WM983B/MRC5) of the key metabolites that distinguish between the MRC5 and WM983B cell lines.

https://doi.org/10.1371/journal.pone.0352639.t002

thumbnail
Download:
Table 3. Fold change of concentrations (HT168-M1/MRC5) of the key metabolites that distinguish between the MRC5 and HT168-M1 cell lines.

https://doi.org/10.1371/journal.pone.0352639.t003

thumbnail
Download:
Table 4. Fold change of concentrations (HT168-M1/WM983B) of the key metabolites that distinguish between the HT168-M1 and WM983B cell lines.

https://doi.org/10.1371/journal.pone.0352639.t004

thumbnail
Download:
Fig 3. Boxplots showing concentrations of the metabolites that distinguish between the MRC5 and WM983B cell lines.

https://doi.org/10.1371/journal.pone.0352639.g003

thumbnail
Download:
Fig 4. Boxplots showing concentrations of the metabolites that distinguish between the MRC5 and HP168-M1 cell lines.

https://doi.org/10.1371/journal.pone.0352639.g004

thumbnail
Download:
Fig 5. Boxplots showing concentrations of the metabolites that distinguish between the HT168-M1 and WM983B cell lines.

https://doi.org/10.1371/journal.pone.0352639.g005

To assess the predictive ability of the PLS-DA models, a 10-fold cross validation with 5 components was performed. In all cases R2 values were above 0.98 and Q2 values were above 0.88, confirming the good predictive ability of the model (MRC5 and WM983B: R2 = 0.994, Q2 = 0.891; MRC5 and HT168-M1: R2 = 0.994, Q2 = 0.895; HT168-M1 and WM983B: R2 = 0.986, Q2 = 0.885)

A comparison of the healthy MRC5 cell line and the WM983B cell line revealed that 16 metabolites were important in discriminating between the two. Higher levels of glutamine, glucose (Gluc) and pyroglutamate (pyroGlu) are seen in MRC5 and all the other metabolites listed in Table 1 are elevated in the melanoma cell line, the most important with high VIP scores being phosphocholine (PCho), aspartate and creatine (Cr) (Fig 3D). Examining the changes in metabolite concentrations between the two cell lines reveals a more than twofold decrease in glucose and glutamine concentrations in the melanoma cell line and a more than twofold increase in UDP-glucose, GTP, creatine, ATP and glutathione concentrations. A similar comparison of the highly metastatic HT168-M1 cell line with MRC5 (Table 3 and Fig 2E) shows higher levels of the same three metabolites (pyroGlu, Gln and Gluc) in MRC5 as in the comparison with the less metastatic cell line. In addition, 1-methylnicotinamide and some amino acids, such as phenylalanine, tryptophan and alanine also have higher levels in MRC5. Besides PCho which has a higher relative concentration in HT168-M1, glutathione (GSH), myo-inositol and glycerophosphocholine (GPC) seem to play the most important role in the discrimination. The highest fold-change can be observed in the concentrations of UDP-glucose, ATP, GPC, 1-methylnicotinamide and glucose.

When the two melanoma cell lines with different metastatic capacities were compared, it was found that GSH, glutamate, UDP-glucose and glutamine were present in higher proportions in the highly metastatic HT168-M1 cell line. In contrast, WM983B had higher levels of creatine phosphate (CrP), serine, taurine, hypotaurine, alanine, β-alanine and asparagine (Table 4 and Fig 2F). The differences in metabolite concentrations between the two melanoma cell lines are much more moderate than those observed in the comparison with healthy cells. Only UDP-glucose, GSH, NAD+, alanine and β-alanine has a fold change greater than 1.5.

Metabolic pathway analysis

To facilitate interpretation of the results, the discriminating metabolite sets (Tables 2–4) were introduced into MetaboAnalyst’s metabolic pathway analysis module. The results of the analysis are summarized in the S5-S7 Tables in S1 File. The topology maps of the altered biochemical pathways are shown in Fig 6, where pathways with a high impact (>0.1) and low p-values (<0.05) are listed.

thumbnail
Download:
Fig 6. Metabolic pathway analysis.

The most affected pathways with high impact (>0.1) and a p-value <0.05 are listed on the right. (A) MRC5 and WM983B cell lines. (B) MRC5 and HT168-M1 cell lines. (C) HT168-M1 and WM983B cell lines.

https://doi.org/10.1371/journal.pone.0352639.g006

When the healthy MRC5 cell line is compared with melanoma cell lines, the most significant altered pathways are starch and sucrose metabolism, alanine, aspartate and glutamate metabolism, and glutathione metabolism. Additionally, the metabolism of β-alanine, pyruvate, arginine and proline are disrupted in WM983B. Meanwhile, the metabolism of aromatic amino acids, nicotinate and nicotinamide are altered in HT168-M1.

The 17 metabolites that were found to differ between the highly metastatic HT168-M1 and the less metastatic WM983B cell lines suggest alterations in the metabolism of taurine and hypotaurine, glutathione and several amino acids.

Discussion

Comparison of the healthy and melanoma cell lines

Examining the NMR spectra of the investigated cell lines (Fig 1), the most notable difference between the MRC5 healthy cell line and the two melanoma cell lines is the presence of multiple high-intensity glucose signals in the MRC5 spectrum in the 3.2–5.3 ppm region. Consistently, PLS-DA analysis identified glucose as a key discriminating metabolite (Table 2 and 3), reflecting the well-known increased glucose consumption of cancer cells [41]. Melanoma cells also exhibitied elevated ATP and lactate levels, indicating enhanced glycolytic activity, and confirming the aerobic glycolytic phenotype of our cell lines [5]. Lactate production acidifies the tumor microenvironment and promotes extracellular matrix degradation, angiogenesis, and immune evasion, all of which are key processes that facilitate invasion and metastatic spread. Furthermore, lactate acts as a signalling molecule that reprograms gene expression and activates pathways (e.g., GPR81, HIF-1α–associated signalling) that drive epithelial–mesenchymal transition, immune suppression, and metastatic colonization [42].

Glutamine and pyroglutamate were more abundant in MRC5 cells. Besides glucose, glutamine is the most important nutrient for cancer cells. It also serves as a source of carbon and nitrogen for the biosynthesis of amino acids, nucleotides, and lipids which are necessary for cell growth and rapid proliferation. Additionally, glutamine plays a key role in maintaining redox homeostasis by promoting the biosynthesis of GSH and NADPH [12,13]. Pyroglutamic acid (also known as 5-oxoproline or pidolic acid) is an intermediate in glutathione metabolism and has been proposed as a potential biomarker for pancreatic and breast cancers [43,44]. As expected, the levels of pyroglutamate in our comparisons are inversely related to those of GSH, an important discriminating metabolite that is upregulated in both WM983B and HT168-M1.

PCho, a key component in membrane synthesis, was also found to be upregulated in our two melanoma cell lines, with a concentration change of approximately 1.5-fold (Table 2 and 3). Elevated levels of PCho were found in most cancer types such as skin-, breast-, ovarian- and prostate cancers [21,45–47], however opposite changes are also reported for a type of lung cancer (NSCLC) [48]. A higher concentration of this metabolite results from increased choline transport into the cells and the overexpression of choline kinase. This contributes to tumor progression and metastasis [49].

There are two amino acids among the metabolites that have elevated levels in both WM983B and HT168-M1, aspartate and ornithine. Aspartate plays a crucial role in the metabolism of cancer cells because it is involved in several vital biochemical processes, such as nucleotide- and protein synthesis, and redox homeostasis, all of which are essential for cells that are proliferating rapidly [50]. Because tumor cells often have difficulty taking up aspartate from their environment, they rely on the mitochondrial catabolism of glutamine to obtain aspartate. Ornithine, a non-proteogenic α-amino acid is mainly used in cancer cells as a substrate for ornithine decarboxylase, a key enzyme in polyamine synthesis [51]. These polyamines such as spermidine and spermine are essential for cancer cell growth and division. Furthermore, elevated ornithine may reflect broader reprogramming of arginin metabolism and urea-cycle associated pathways, which are increasingly recognised as contributors to tumor growth and microenvironmental adaptation [52].

Finally, UDP-glucose (UDPG) is also another metabolite that has been found to be different between healthy and melanoma cell lines, with higher levels in the latter. UDPG plays a role in regulating cell migration, proliferation, and extracellular matrix modulation. UDPG is converted to UDP-glucuronic acid, a precursor for proteoglycans and glycosaminoglycans, by UDPG dehydrogenase, an enzyme that is overexpressed in several cancers [53].

In the pairwise comparisons between healthy and melanoma cell lines, a few metabolites were found to be able to discriminate between healthy and only one melanoma cell lines. In the pathway analysis comparing MRC5 and WM983B, β-alanine metabolism has the second highest impact (Fig 6A). β-Alanine is a non-essential amino acid and the precursor of carnitine, which plays an important role in energy production and fatty acid metabolism [54]. Kou et al. reported a significantly higher intracellular concentration of β-alanine in aggressive breast cancer cells than in non-cancerous cells [26].

Pyruvate metabolism and arginine and proline metabolism are also altered in the less metastatic cell line. The former includes two metabolites identified in our analysis: pyruvate and lactate. The higher levels of these metabolites in the melanoma cell line suggest enhanced glycolytic flux and/or reduced mitochondrial oxidative utilization of pyruvate, reflecting reprogrammed energy metabolism [55]. Pyruvate is also part of the arginine and proline metabolism pathway, along with ornithine and creatine. This latter metabolite plays an important role in the energy buffering system by forming phosphocreatine, which can rapidly generate ATP when needed, by the enzyme phosphocreatine-creatine kinase (CK). Given that rapid cell division requires immediate energy, it is not surprising that abnormal CK levels are a feature of several cancers [56].

Comparing the metabolites of the healthy cell line with those of the highly metastatic HT168-M1, unique disturbances are observed in the biosynthesis of the aromatic amino acids (phenylalanine, tyrosine and tryptophan), as well as in the metabolism of phenylalanine, nicotinate, and nicotinamide (Fig 6B). Phenylalanine and tryptophan levels were lower in the highly metastatic cell line. These two essential amino acids are crucial for the protein synthesis, with phenylalanine acting as a precursor to tyrosine, which in turn is a precursor to melanin, a pigment produced in melanocytes [57]. The metabolism of phenylalanine is often altered in cancer cells, particularly affecting the mTORC1 signalling pathway. This pathway is a critical regulator of cell growth, proliferation and lipid metabolism [58]. Tryptophan is an essential amino acid that plays a key role in cancer metabolism, primarily via the kynurenine pathway. In tumors of the digestive system, lower levels of tryptophan have been found to correlate with increased tumour cell migration, immune suppression and increased metastatic potential [59,60]. Altered amino acid metabolism in cancer cells can have a profound effect on the tumor microenvironment, suppressing or reprogramming immune cell function, particularly that of memory T-cells [61]. 1-Methylnicotinamide (MNA) is a metabolite of the nicotinate and nicotinamide metabolic pathway synthesized from nicotinamide via N-methylation by the nicotinamide-N-methyltransferase enzyme, which has been found to be overexpressed in several human cancers [32,62]. MNA is an immune-regulatory metabolite in human ovarian cancer that induces T-cells to secrete tumour necrosis factor alpha [63].

There are three other metabolites in Table 3 to be mentioned. Glycerophosphocholine is a choline metabolite, and its level is higher in HT168-M1, similarly to that of PCho. The glucose-derived metabolite, myo-inositol, is a precursor of phosphoinositides, which are recognized as signalling molecules in cells. The PI3K and PTEN enzymes, which regulate the level of phosphatidylinositol-3,4,5-triphosphate, are frequently mutated in cancer [64]. Finally, the non-essential amino acid alanine, can be produced directly from pyruvate by the enzyme alanine transaminase. It is used in protein synthesis and is also a glycolytic by-product that transports excess carbon from glycolysis [65]. Intensive use of this amino acid results in low levels in HT168-M1 cells.

Comparison of the melanoma cell lines with high and low metastatic capacity

When evaluating the results obtained from comparing the metabolites from the two melanoma cell lines, two pathways were found to be important in all the pairwise comparisons: glutathione metabolism, and alanine, aspartate and glutamate metabolism (Fig 6). The former includes higher levels of GSH and glutamate, and lower levels of pyroglutamate in the highly metastatic cell line (Table 4). This highlights the importance of maintaining high antioxidant levels in the highly metastatic cell line, in order to cope with the reactive oxygen species produced by intense metabolic activity. These metabolic adaptations may be linked to broader pathway-regulated transcriptional programs, since AP-1-associated signaling enhance cellular antioxidant defense by upregulating genes involved in glutathione synthesis, and also in amino acid transport and metabolism [66]. An earlier study by Carretero et al. demonstrated that B16 melanoma cells with a higher GSH content exhibited greater metastatic activity in the liver than cells with a low GSH content [67]. Liver-metastatic breast cancer cells also exhibit upregulated GSH metabolism [68]. GSH has importance in key signal transduction reactions as a controller of cell differentiation, proliferation, apoptosis, ferroptosis and immune function [69]. As glutamate is a precursor of GSH, an elevated level of glutamate also supports GSH synthesis and, consequently, the oxidative stress resistance of cancer cells. In metastatic renal cell carcinoma and hepatocellular carcinoma, elevated glutamate levels have been associated with increased glutaminase activity [27,70].

The metabolites of the alanine, aspartate and glutamate metabolism pathway that distinguish the two melanoma cell lines are: alanine, pyruvate, asparagine, glutamate and glutamine. Due to the high energy demand and intensive protein synthesis, the levels of both alanine and pyruvate are lower in the highly metastatic cell line than in the low metastatic cell line. Asparagine is essential for cancer growth and development, and is a necessary component of the proteins involved in the epithelial-mesenchymal transition which promotes metastatic processes [71]. The lower level of asparagine in the HT168-M1 cell line indicates more intense metabolism of this amino acid than in the less metastatic cell line. Glutamine was found to discriminate between healthy and melanoma cell lines, as well as between melanoma cell lines with high and low metastatic capacity. The higher level of glutamine in HT168-M1 suggests that this cell line relies less on glutamine for energy production.

The pathway analysis comparing the two melanoma cell lines revealed that taurine and hypotaurine metabolism had the greatest impact, despite these metabolites having a lower fold change in concentration. WM983B has higher levels of taurine and hypotaurine relative to MRC5. Hypotaurine is an intermediate in the biosynthesis of taurine and produced by the enzyme cysteamine dioxygenase, which plays an important role in protecting against oxidative stress. Earlier studies have shown that this metabolic pathway is disrupted in several types of cancer [72–74]. Gao et al. found a strong positive correlation between glioma grade and hypotaurine levels, and demonstrated that hypotaurine activates hypoxia signaling [75]. Taurine, the oxidation product of hypotaurine, can act as a direct antioxidant by reducing the reactive oxygen species and increasing the activity of certain antioxidant enzymes [76,77].

In all the pairwise analyses, UDP-glucose is a discriminating metabolite, with levels increasing from the healthy cell line to the highly metastatic HT168-M1 cell line. In addition to the metabolites discussed so far, the level of some amino acids such as serine, isoleucine and histidine differ in the two melanoma cell lines, having higher levels in the less metastatic WM983B.

Creatine phosphate is an important metabolite that distinguishes the two melanoma cell lines from each other, as it stores energy that can be released quickly. The highly metastatic cell line has lower levels of creatine phosphate due to its intensive energy use. Levels of NAD+ are also low in HT168-M1. It is an essential metabolite in oxidation-reduction reactions during energy production [78]. Alterations in NAD+ levels may reflect shifts in cellular redox balance and mitochondrial metabolism; however, the concurrent lower pyruvate and higher GSH levels suggest redox imbalance. 3-Methyl-2-oxovalerate is the product of incomplete breakdown of branched-chain amino acids, and its level is elevated in the low metastatic cell line.

The metabolic biomarkers identified in this study may also have translational relevance, as emerging nanoparticle-based precision oncology approaches increasingly aim to exploit tumor-selective metabolic vulnerabilities [79].

Previous studies of melanoma cells derived from the primary (WM-115) and metastatic (WM-226-4) sites of the same patient identified glycerophospholipid metabolism as the most affected pathway, followed by the carnitine shuttle, which has a biological implication in the survival of metastatic cells. However, it was also found that tryptophan and branched-chain amino acid metabolism were altered [80]. Kosmopoulou et al. investigated the same two cell lines, highlighting the critical roles of purine, pyrimidine and amino acid metabolism in human melanoma metastasis [30]. Another study of melanoma cells with different metastatic potential revealed an accumulation of organic acids, particularly lactic acid, in melanoma cell lines compared to melanocytes. This accumulation increased progressively with increasing metastatic potential. Additionally, elevated levels of glucose-6-phosphate and several amino acids, including glycine, pyroglutamic acid, cysteine, ornithine, tyrosine and lysine, were found in the highly metastatic cell lines. However, the most significant affected pathway was alanine, aspartate and glutamate metabolism [14].

The metabolic pathway alterations identified in our study are in agreement with earlier findings. The apparent differences can be attributed to the different cell lines tested or the different analytical techniques employed.

Conclusion

Our metabolomic analysis revealed distinct metabolic reprogramming between healthy fibroblasts and melanoma cell lines, as well as between melanoma cells of low and high metastatic potential. The findings highlight enhanced glycolysis and glutathione-dependent antioxidant defense as key features of melanoma metabolism, with the highly metastatic cells displaying stronger redox adaptation and altered amino acid utilization. Elevated glutathione and glutamate, together with reduced NAD⁺ and pyruvate levels, indicate a metabolic shift favoring oxidative stress resistance and survival under high energy demand. However, as this study was limited by the use of only two melanoma cell lines and in vitro conditions, future research integrating larger cell cohorts, multi-omics approaches, and in vivo validation will be essential to confirm and extend these findings.

Supporting information

S1 File. Supporting Figures and Tables.

Assignation of metabolites on the 1H-NMR spectrum of a HT168M1 cell extract (S1 Fig.); The 1H-1H TOCSY NMR (600 MHz) spectrum of the melanoma cell line extracts (S2-S3 Figs); PCA scores plots of the cell lines studied using pairwise analysis (S4 Fig.); PLS-DA scores plots of the three cell lines studied (S5 Fig.); VIP scores by component 1 in the PLS-DA analysis of the investigated cell lines (S1-S3 Tables); Selective integration regions of the most important metabolites (S4 Table); Pathway analysis results (S5-S7 Tables).

https://doi.org/10.1371/journal.pone.0352639.s001

(PDF)

S2 File. Binned NMR data of the studied cell extracts.

https://doi.org/10.1371/journal.pone.0352639.s002

(XLSX)

Acknowledgments

János Bolyai Research Scholarship of the Hungarian Academy of Sciences is appreciated by A.M.

References

  1. 1. Gosman LM, Țăpoi D-A, Costache M. Cutaneous Melanoma: A Review of Multifactorial Pathogenesis, Immunohistochemistry, and Emerging Biomarkers for Early Detection and Management. Int J Mol Sci. 2023;24(21):15881. pmid:37958863
  2. 2. Siegel RL, Kratzer TB, Wagle NS, Sung H, Jemal A. Cancer statistics, 2026. CA Cancer J Clin. 2026.
  3. 3. Pote MS, Singh D, M A A, Suchita J, Gacche RN. Cancer metastases: Tailoring the targets. Heliyon. 2024;10(15):e35369. pmid:39170575
  4. 4. Liberti MV, Locasale JW. The Warburg Effect: How Does it Benefit Cancer Cells?. Trends Biochem Sci. 2016;41(3):211–8. pmid:26778478
  5. 5. Lunt SY, Vander Heiden MG. Aerobic glycolysis: meeting the metabolic requirements of cell proliferation. Annu Rev Cell Dev Biol. 2011;27:441–64.
  6. 6. Bergers G, Fendt S-M. The metabolism of cancer cells during metastasis. Nat Rev Cancer. 2021;21(3):162–80. pmid:33462499
  7. 7. Ruocco MR, Avagliano A, Granato G, Vigliar E, Masone S, Montagnani S, et al. Metabolic flexibility in melanoma: A potential therapeutic target. Semin Cancer Biol. 2019;59:187–207. pmid:31362075
  8. 8. Zhang J, Yao Y-H, Li B-G, Yang Q, Zhang P-Y, Wang H-T. Prognostic value of pretreatment serum lactate dehydrogenase level in patients with solid tumors: a systematic review and meta-analysis. Sci Rep. 2015;5:9800. pmid:25902419
  9. 9. Fischer GM, Carapeto FCL, Joon AY, Haydu LE, Chen H, Wang F, et al. Molecular and immunological associations of elevated serum lactate dehydrogenase in metastatic melanoma patients: A fresh look at an old biomarker. Cancer Med. 2020;9(22):8650–61. pmid:33016647
  10. 10. Costantini S, Madonna G, Capone M, Di Gennaro E, Bagnara P, Renza F, et al. Metabolomic signatures in liquid biopsy are associated with overall survival in metastatic melanoma patients treated with immune checkpoint inhibitor therapy. J Exp Clin Cancer Res. 2025;44(1):119. pmid:40211360
  11. 11. Rodrigues MF, Obre E, de Melo FHM, Santos GC Jr, Galina A, Jasiulionis MG, et al. Enhanced OXPHOS, glutaminolysis and β-oxidation constitute the metastatic phenotype of melanoma cells. Biochem J. 2016;473(6):703–15. pmid:26699902
  12. 12. Cluntun AA, Lukey MJ, Cerione RA, Locasale JW. Glutamine Metabolism in Cancer: Understanding the Heterogeneity. Trends Cancer. 2017;3(3):169–80. pmid:28393116
  13. 13. Gong T, Zheng C, Ou X, Zheng J, Yu J, Chen S, et al. Glutamine metabolism in cancers: Targeting the oxidative homeostasis. Front Oncol. 2022;12:994672. pmid:36324588
  14. 14. Kim H-Y, Lee H, Kim S-H, Jin H, Bae J, Choi H-K. Discovery of potential biomarkers in human melanoma cells with different metastatic potential by metabolic and lipidomic profiling. Sci Rep. 2017;7(1):8864. pmid:28821754
  15. 15. Qi S-A, Wu Q, Chen Z, Zhang W, Zhou Y, Mao K, et al. High-resolution metabolomic biomarkers for lung cancer diagnosis and prognosis. Sci Rep. 2021;11(1):11805. pmid:34083687
  16. 16. Zhang X, Xia B, Zheng H, Ning J, Zhu Y, Shao X, et al. Identification of characteristic metabolic panels for different stages of prostate cancer by 1H NMR-based metabolomics analysis. J Transl Med. 2022;20(1):275. pmid:35715864
  17. 17. Cao K, Lyu Y, Chen J, He C, Lyu X, Zhang Y, et al. Prognostic Implication of Plasma Metabolites in Gastric Cancer. Int J Mol Sci. 2023;24(16):12774. pmid:37628957
  18. 18. Kaur R, Gupta S, Kulshrestha S, Khandelwal V, Pandey S, Kumar A, et al. Metabolomics-Driven Biomarker Discovery for Breast Cancer Prognosis and Diagnosis. Cells. 2024;14(1):5. pmid:39791706
  19. 19. Silva C, Perestrelo R, Silva P, Tomás H, Câmara JS. Breast Cancer Metabolomics: From Analytical Platforms to Multivariate Data Analysis. A Review. Metabolites. 2019;9(5):102. pmid:31121909
  20. 20. Li T, Deng P. Nuclear magnetic resonance technique in tumor metabolism. Genes Dis. 2017;4:28–36.
  21. 21. Hagyousif YA, Sharaf BM, Zenati RA, El-Huneidi W, Bustanji Y, Abu-Gharbieh E, et al. Skin Cancer Metabolic Profile Assessed by Different Analytical Platforms. Int J Mol Sci. 2023;24(2):1604. pmid:36675128
  22. 22. Lou Y, Cheng M, Cao Q, Li K, Qin H, Bao M, et al. Simultaneous quantification of mirabegron and vibegron in human plasma by HPLC-MS/MS and its application in the clinical determination in patients with tumors associated with overactive bladder. J Pharm Biomed Anal. 2024;240:115937. pmid:38198885
  23. 23. Wu L, Zhang Y, Zhang X, Xiong X, Liu W. Development, validation and clinical application of a LC-MS/MS method for simultaneous determination of third-generation EGFR-TKIs and metabolites in patients with NSCLC. Microchem J. 2026;224:17710.
  24. 24. Emwas A-H, Roy R, McKay RT, Tenori L, Saccenti E, Gowda GAN, et al. NMR Spectroscopy for Metabolomics Research. Metabolites. 2019;9(7):123. pmid:31252628
  25. 25. Debik J, Sangermani M, Wang F, Madssen TS, Giskeødegård GF. Multivariate analysis of NMR-based metabolomic data. NMR Biomed. 2022;35(2):e4638. pmid:34738674
  26. 26. Kou F, Zhu B, Zhou W, Lv C, Cheng Y, Wei H. Targeted metabolomics reveals dynamic portrayal of amino acids and derivatives in triple-negative breast cancer cells and culture media. Mol Omics. 2021;17(1):142–52. pmid:33295912
  27. 27. Yang Y, Li C, Nie X, Feng X, Chen W, Yue Y, et al. Metabonomic studies of human hepatocellular carcinoma using high-resolution magic-angle spinning 1H NMR spectroscopy in conjunction with multivariate data analysis. J Proteome Res. 2007;6(7):2605–14. pmid:17564425
  28. 28. McCaslin L, Lagies S, Mohl DA, Plattner DA, Jäger M, Nöthling C, et al. Unveiling Metabolic Subtypes in Endometrial Cancer Cell Lines: Insights from Metabolomic Analysis Under Standard and Stress Conditions. Int J Mol Sci. 2025;26(19):9573. pmid:41096838
  29. 29. Silva CL, Perestrelo R, Silva P, Tomás H, Câmara JS. Volatile metabolomic signature of human breast cancer cell lines. Sci Rep. 2017;7:43969. pmid:28256598
  30. 30. Kosmopoulou M, Giannopoulou AF, Iliou A, Benaki D, Panagiotakis A, Velentzas AD, et al. Human Melanoma-Cell Metabolic Profiling: Identification of Novel Biomarkers Indicating Metastasis. Int J Mol Sci. 2020;21(7):2436. pmid:32244549
  31. 31. Jiang Z, Chen Z, Xu Y, Li H, Li Y, Peng L, et al. Low-Frequency Ultrasound Sensitive Piezo1 Channels Regulate Keloid-Related Characteristics of Fibroblasts. Adv Sci (Weinh). 2024;11(14):e2305489. pmid:38311578
  32. 32. Eckert MA, Coscia F, Chryplewicz A, Chang JW, Hernandez KM, Pan S, et al. Proteomics reveals NNMT as a master metabolic regulator of cancer-associated fibroblasts. Nature. 2019;569(7758):723–8. pmid:31043742
  33. 33. Hoek KS, Schlegel NC, Brafford P, Sucker A, Ugurel S, Kumar R, et al. Metastatic potential of melanomas defined by specific gene expression profiles with no BRAF signature. Pigment Cell Res. 2006;19(4):290–302. pmid:16827748
  34. 34. Kenessey I, Keszthelyi M, Krámer Z, Berta J, Adám A, Dobos J, et al. Inhibition of c-Met with the specific small molecule tyrosine kinase inhibitor SU11274 decreases growth and metastasis formation of experimental human melanoma. Curr Cancer Drug Targets. 2010;10(3):332–42. pmid:20370683
  35. 35. Ladányi A, Tímár J, Paku S, Molnár G, Lapis K. Selection and characterization of human melanoma lines with different liver-colonizing capacity. Int J Cancer. 1990;46(3):456–61. pmid:2203689
  36. 36. Kenessey I, Kramer Z, István L, Cserepes MT, Garay T, Hegedűs B, et al. Inhibition of epidermal growth factor receptor improves antitumor efficacy of vemurafenib in BRAF-mutant human melanoma in preclinical model. Melanoma Res. 2018;28(6):536–46. pmid:30124539
  37. 37. Pang Z, Lu Y, Zhou G, Hui F, Xu L, Viau C, et al. MetaboAnalyst 6.0: towards a unified platform for metabolomics data processing, analysis and interpretation. Nucleic Acids Res. 2024;52(W1):W398–406. pmid:38587201
  38. 38. Chong J, Wishart DS, Xia J. Using MetaboAnalyst 4.0 for Comprehensive and Integrative Metabolomics Data Analysis. Curr Protoc Bioinformatics. 2019;68(1):e86. pmid:31756036
  39. 39. Kanehisa M, Furumichi M, Sato Y, Kawashima M, Ishiguro-Watanabe M. KEGG for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res. 2023;51(D1):D587–92. pmid:36300620
  40. 40. Wishart DS, Guo A, Oler E, Wang F, Anjum A, Peters H, et al. HMDB 5.0: the Human Metabolome Database for 2022. Nucleic Acids Res. 2022;50(D1):D622–31. pmid:34986597
  41. 41. Fadaka A, Ajiboye B, Ojo O, Adewale O, Olayide I, Emuowhochere R. Biology of glucose metabolization in cancer cells. J Oncol Sci. 2017;3:45–51.
  42. 42. Liu H, Pan M, Liu M, Zeng L, Li Y, Huang Z, et al. Lactate: a rising star in tumors and inflammation. Front Immunol. 2024;15:1496390. pmid:39660139
  43. 43. Wang W, Tian X, Yan L, Guan X, Dong B, Zhao M, et al. Identification of the γ-glutamyl cycle as a novel therapeutic target and 5-oxoproline as a new biomarker for diagnosing pancreatic cancer. Ann Med. 2023;55(2):2242247. pmid:37544888
  44. 44. Park J, Shin Y, Kim TH, Kim D-H, Lee A. Plasma metabolites as possible biomarkers for diagnosis of breast cancer. PLoS One. 2019;14(12):e0225129. pmid:31794572
  45. 45. Eliyahu G, Kreizman T, Degani H. Phosphocholine as a biomarker of breast cancer: molecular and biochemical studies. Int J Cancer. 2007;120(8):1721–30. pmid:17236204
  46. 46. Bagnoli M, Granata A, Nicoletti R, Krishnamachary B, Bhujwalla ZM, Canese R, et al. Choline Metabolism Alteration: A Focus on Ovarian Cancer. Front Oncol. 2016;6:153. pmid:27446799
  47. 47. Kdadra M, Höckner S, Leung H, Kremer W, Schiffer E. Metabolomics Biomarkers of Prostate Cancer: A Systematic Review. Diagnostics (Basel). 2019;9(1):21. pmid:30791464
  48. 48. Hu J-M, Sun H-T. Serum proton NMR metabolomics analysis of human lung cancer following microwave ablation. Radiat Oncol. 2018;13(1):40. pmid:29530051
  49. 49. Jin H-R, Wang J, Wang Z-J, Xi M-J, Xia B-H, Deng K, et al. Lipid metabolic reprogramming in tumor microenvironment: from mechanisms to therapeutics. J Hematol Oncol. 2023;16(1):103. pmid:37700339
  50. 50. Gorgoglione R, Impedovo V, Riley CL, Fratantonio D, Tiziani S, Palmieri L, et al. Glutamine-Derived Aspartate Biosynthesis in Cancer Cells: Role of Mitochondrial Transporters and New Therapeutic Perspectives. Cancers (Basel). 2022;14(1):245. pmid:35008407
  51. 51. Wang H, Qin K, Shi D, Wu P, Hao X, Liu H, et al. A new 68Ga-labeled ornithine derivative for PET imaging of ornithine metabolism in tumors. Amino Acids. 2023;55(5):595–606. pmid:36809562
  52. 52. Sivashanmugam M, J J, V U, K N S. Ornithine and its role in metabolic diseases: An appraisal. Biomed Pharmacother. 2017;86:185–94. pmid:27978498
  53. 53. Price MJ, Nguyen AD, Byemerwa JK, Flowers J, Baëta CD, Goodwin CR. UDP-glucose dehydrogenase (UGDH) in clinical oncology and cancer biology. Oncotarget. 2023;14:843–57. pmid:37769033
  54. 54. Flanagan JL, Simmons PA, Vehige J, Willcox MD, Garrett Q. Role of carnitine in disease. Nutr Metab (Lond). 2010;7:30. pmid:20398344
  55. 55. Shen D, Zhang L, Li S, Tang L. Metabolic reprogramming in melanoma therapy. Cell Death Discov. 2025;11(1):308. pmid:40617808
  56. 56. Yan Y-B. Creatine kinase in cell cycle regulation and cancer. Amino Acids. 2016;48(8):1775–84. pmid:27020776
  57. 57. Vargas AJ, Sittadjody S, Thangasamy T, Mendoza EE, Limesand KH, Burd R. Exploiting tyrosinase expression and activity in melanocytic tumors: quercetin and the central role of p53. Integr Cancer Ther. 2011;10(4):328–40. pmid:21196432
  58. 58. Bagheri P, Hoang K, Fung AA, Hussain S, Shi L. Visualizing Cancer Cell Metabolic Dynamics Regulated With Aromatic Amino Acids Using DO-SRS and 2PEF Microscopy. Front Mol Biosci. 2021;8:779702. pmid:34977157
  59. 59. Wiggins T, Kumar S, Markar SR, Antonowicz S, Hanna GB. Tyrosine, phenylalanine, and tryptophan in gastroesophageal malignancy: a systematic review. Cancer Epidemiol Biomarkers Prev. 2015;24(1):32–8. pmid:25344892
  60. 60. Yu L, Lu J, Du W. Tryptophan metabolism in digestive system tumors: unraveling the pathways and implications. Cell Commun Signal. 2024;22(1):174. pmid:38462620
  61. 61. Sun D-Y, Hu Y-J, Li X, Peng J, Dai Z-J, Wang S. Unlocking the full potential of memory T cells in adoptive T cell therapy for hematologic malignancies. Int Immunopharmacol. 2025;144:113392. pmid:39608170
  62. 62. Ulanovskaya OA, Zuhl AM, Cravatt BF. NNMT promotes epigenetic remodeling in cancer by creating a metabolic methylation sink. Nat Chem Biol. 2013;9(5):300–6. pmid:23455543
  63. 63. Kilgour MK, MacPherson S, Zacharias LG, Ellis AE, Sheldon RD, Liu EY, et al. 1-Methylnicotinamide is an immune regulatory metabolite in human ovarian cancer. Sci Adv. 2021;7(4):eabe1174. pmid:33523930
  64. 64. Bunney TD, Katan M. Phosphoinositide signalling in cancer: beyond PI3K and PTEN. Nat Rev Cancer. 2010;10(5):342–52. pmid:20414202
  65. 65. DeBerardinis RJ, Cheng T. Q’s next: the diverse functions of glutamine in metabolism, cell biology and cancer. Oncogene. 2010;29(3):313–24. pmid:19881548
  66. 66. Li F, Tian J, Zhang L, He H, Song D. A multi-omics approach to reveal critical mechanisms of activator protein 1 (AP-1). Biomed Pharmacother. 2024;178:117225. pmid:39084078
  67. 67. Carretero J, Obrador E, Anasagasti MJ, Martin JJ, Vidal-Vanaclocha F, Estrela JM. Growth-associated changes in glutathione content correlate with liver metastatic activity of B16 melanoma cells. Clin Exp Metastasis. 1999;17(7):567–74. pmid:10845555
  68. 68. Biondini M, Lehuédé C, Tabariès S, Annis MG, Pacis A, Ma EH, et al. Metastatic breast cancer cells are metabolically reprogrammed to maintain redox homeostasis during metastasis. Redox Biol. 2024;75:103276. pmid:39053265
  69. 69. Kennedy L, Sandhu JK, Harper M-E, Cuperlovic-Culf M. Role of Glutathione in Cancer: From Mechanisms to Therapies. Biomolecules. 2020;10(10):1429. pmid:33050144
  70. 70. Gao H, Dong B, Jia J, Zhu H, Diao C, Yan Z, et al. Application of ex vivo (1)H NMR metabonomics to the characterization and possible detection of renal cell carcinoma metastases. J Cancer Res Clin Oncol. 2012;138(5):753–61. pmid:22258851
  71. 71. Yuan Q, Yin L, He J, Zeng Q, Liang Y, Shen Y, et al. Metabolism of asparagine in the physiological state and cancer. Cell Commun Signal. 2024;22(1):163. pmid:38448969
  72. 72. Huang S, Chong N, Lewis NE, Jia W, Xie G, Garmire LX. Novel personalized pathway-based metabolomics models reveal key metabolic pathways for breast cancer diagnosis. Genome Med. 2016;8(1):34. pmid:27036109
  73. 73. Fong MY, McDunn J, Kakar SS. Identification of metabolites in the normal ovary and their transformation in primary and metastatic ovarian cancer. PLoS One. 2011;6(5):e19963. pmid:21625518
  74. 74. Samaržija I, Trošelj KG, Konjevoda P. Prognostic Significance of Amino Acid Metabolism-Related Genes in Prostate Cancer Retrieved by Machine Learning. Cancers (Basel). 2023;15(4):1309. pmid:36831650
  75. 75. Gao P, Yang C, Nesvick CL, Feldman MJ, Sizdahkhani S, Liu H, et al. Hypotaurine evokes a malignant phenotype in glioma through aberrant hypoxic signaling. Oncotarget. 2016;7(12):15200–14. pmid:26934654
  76. 76. Anand P, Rajakumar D, Jeraud M, Felix AJW, Balasubramanian T. Effects of taurine on glutathione peroxidase, glutathione reductase and reduced glutathione levels in rats. Pak J Biol Sci. 2011;14(3):219–25. pmid:21870645
  77. 77. Vohra BP, Hui X. Taurine protects against carbon tetrachloride toxicity in the cultured neurons and in vivo. Arch Physiol Biochem. 2001;109(1):90–4. pmid:11471076
  78. 78. Yong J, Cai S, Zeng Z. Targeting NAD metabolism: dual roles in cancer treatment. Front Immunol. 2023;14:1269896.
  79. 79. Altaf S, Fatima Q, Bisma, Ramzan A, Sarwar M, Rani B, et al. Tailored nanoparticles in precision animal and human oncology: engineering tumor-selective delivery and microenvironment interplay. Agrobiol Records. 2025;22:61–72.
  80. 80. Yu Z, Huang M, Clowers BH. Comparative metabolite profiling of a metastatic and primary melanoma cell line using untargeted metabolomics: A case study. Clin Mass Spectrom. 2018;10:16–24. pmid:39193356