Table 2.

Summary of transcriptomic–metabolomic integration studies for PCa within the last decade (2011–2021) ^1,2.

Reference	Experimental Condition	Sample/ n Samples	Analytical Tool for Metabolites	Altered Metabolites (+/−)	Dysregulated Metabolic Pathways	Main Findings
Imir et al., 2021 [147]	Perfluoroalkyl sulfonate (PFAS) exposure	RWPE-1 RWPE-kRAS	GC-MS	Acetyl-coA Pyruvate dehydrogenase complex (PDC)	Glycolysis via Warburg effect and transfer of acetyl group into mitochondria TCA cycle Threonine and 2-oxobutanoate degradation Phosphatidylethanol-amine biosynthesis Lysine degradation Pentose phosphate pathway (PPP)	PFAS exposure led to increase in xenograft tumor growth and altered metabolic phenotype of PCa, particularly those associated w/ glucose metabolism via the Warburg effect, involving the transfer of acetyl groups into mitochondria and TCA (pyruvate). PFAS increased PPAR signaling and histone acetylation in PCa.
Tilborg and Saccenti 2021 [224]	Gene expression-metabolic dysregulation relationships	14 metabolic data sets, one of those is for PCa. 7 = tissue PCa 7 = tissue normal	Statistical, no experimental tool	Out of 72 metabolites investigated in PCa, 0 significantly differentially abundant metabolites were found (padj < 0.05)	No enriched or dysregulated pathways for PCa	Topological analysis of Gaussian networks → PCa more defined by genetic networks than metabolic ones. PCa-related metabolites were not significantly altered between controls and PCa samples.
Wang et al., 2021 [225]	Differential metabolites between PCa and BHP	41 = PCa 38 = BPH	GC-MS GC/Q-TOF-MS Multivariate and univariate statistical analysis	12 metabolites (+/−) including L-serine, myo-inositol, and decanoic acid	L-serine, myo-inositol, and decanoic acid metabolism	L-serine, myo-inositol, and decanoic acid → potential biomarkers for discriminating PCa from BHP. The 3 metabolites → increased area under the curve (AUC) of cPSA and tPSA from 0.542 and 0.592 to 0.781, respectively.
Gómez-Cebrián et al., 2020 [226]	Dysregulated PCa metabolic pathway mapping	73 using serum and urine	NMR	36 metabolites (+/−) including glucose, glycine, 1-methylnicotinamide	Energy metabolism Nucleotide synthesis	36 metabolic pathways were dysregulated in PCa based on Gleason score (GS) (low-GS (GS < 7), high-GS PCa (GS ≥ 7) groups). Levels of glucose, glycine, and 1-methylnicotinamide → significantly altered between Gleason groups.
Chen et al., 2020 [148]	EMT-PCa and epithelial PCa differentiation	ARCaPE ARCaPM	LC-MS Glucose uptake assay	Aspartate (+) Glycolytic enzymes (+) except for glucose 2 transporter (−) TCA cycle: pyruvate dehydrogenase kinase 1/2, pyruvate dehydrogenase 2 (+) Succinate dehydrogenase A, aconitase 2 (−) Glutaminase 1/2 (+)	Glucose uptake Aspartate metabolism Glycolysis TCA cycle Glutamine–glutamate conversion	PCa cells undergoing epithelial-mesenchymal transition (EMT) showed low glucose consumption. Glucose metabolism in ARCaPE downregulated. Glucose metabolism in transcription factor- (TF) induced EMT models downregulated. ARCaPM cells showed increased aspartate metabolism.
Joshi et al., 2020 [149]	Carnitine palmitoyl transferase I (CPT1A) expression	LNCaP-C4-2	UPHLC-MS	Acyl-carnitines Mitochondrial reactive oxygen species Superoxide dismutase 2	ER stress Serine biosynthesis Lipid catabolism Androgen response	Upregulated pathways via transcriptomic analysis → ER stress, serine biosynthesis, lipid catabolism. Overexpressed (OE) of CPT1A showed increased SOD2 when subjected to low fatty acids and no androgen → better antioxidant defense w/ CPT1A OE. High lipid metabolism, low androgen response → worse progression-free survival.
Lee et al., 2020 [162]	Urine-enriched mRNA characteriza-tion	Urine: 20 = BPH 11 = PTT 20 = PCa 20 = normal 65 = PCa (validation)	UHPLC-HRMS	Alanine, aspartate, and glutamate (+) Glutamic-oxaloacetic transaminase 1 (+)	14 metabolic pathways including aminoacyl-tRNA biosynthesis TCA cycle Pyruvate metabolism Amino acid pathways	Integrated gene expression-metabolite signature analysis → glutamate metabolism and TCA aberration contributed to PCa phenotype via GOT1-mediated redox balance.
Marin de Mas et al., 2019 [150]	Aldrin exposure analysis via gene-protein-reactions (GPR) associations	DU145	Dataset processing, no experimental tool	19 metabolites, both consuming and producing	Carnitine shuttle Prostaglandin biosynthesis	The application of novel stoichiometric gene–protein reaction (S-GPR) (imbedded in genome-scale metabolic models, GSMM) on the transcriptomic data of Aldrin-exposed DU145 PCa revealed increased metabolite use/production. Carnitine shuttle and prostaglandin biosynthesis → significantly altered in Aldrin-exposed DU145 PCa.
Andersen et al., 2018 [227]	Differential genes and metabolites	158 tissue samples from 43 patients	HR-MAS MRS	23 metabolites differentially expressed between high RSG and low RSG, including spermine, taurine, scyllo-inositol, and citrate	Immunity and ECM remodeling DNA repair pathway Type I interferon signaling	High RSG (≥16%) was associated w/ PCa biochemical recurrence (BCR). These high reactive stromata → upregulated genes and metabolites involved in immune functions and ECM remodeling.
Shao et al., 2018 [228]	Metabolomics-RNA-seq analysis	Tissue: 21 = PCa 21 = normal 50 = PCa and normal each (validation)	GC-MS	Fumarate Malate Branched-chain amino acid (+) Glutaminase, glutamate dehydrogenase ½ (+) Pyruvate dehydrogenase (+)	TCA cycle BCAA degradation Glutamine catabolism Pyruvate catabolism	Fumarate and malate levels → highly correlated w/ Gleason score, tumor stage, and expression of genes involved in BCAA degradation. BCAA degradation, glutamine catabolism, and pyruvate catabolism replenished TCA cycle metabolites.
Al Khadi et al., 2017 [229]	Peripheral and transitional zone differentiation	20 PCa patients undergoing prostatectomy	Network-based integrative analysis, no experimental tool	23 metabolites (+) including fatty acid synthase (FC = 2.9) and ELOVL fatty acid elongase 2 (FC = 2.8)	15 KEGG pathways including de novo lipogenesis and fatty acid β-oxidation	RNA sequencing and high-throughput metabolic analyses (non-cancerous tissue, prostatectomy patients) → genes involved in de novo lipogenesis: peripheral > transitional. Peripheral zone induced lipo-rich priming → PCa oncogenesis.
Sandsmark et al., 2017 [230]	CWP, NCWP, EMT evaluation	129 1519 samples (validation)	HR-MAS MRS MRSI	Citrate (−) Spermine (−)	TCA cycle	Increased NCWP activation via Wnt5a/Fzd2 Wnt activation mode → common in PCa. NCWP activation is associated w/ high EMT expression and high Gleason score. NCWP-EMT → significant predictor of PCa metastasis and biochemical recurrence.
Ren et al., 2016 [231]	Paired approach for altered pathways determination	25 = PCa and adjacent non-cancerous tissues each 51 = PCa and 16 = BHP (validation)	LC-MS TOF-MS	Sphingosine (+) Sphingosine-1-phosphate receptor 2 (−) Choline, S-adenosylhomoserine, 5- methylthioadensine, S-adenosylmethionine, Nicotinamide mononucleotide, Nicotinamide adenine dinucleotide, and Nicotinamide adenine dinucleotide phosphate (+) Adenosine, uric acid (−)	Cysteine metabolism Methionine metabolism Nicotinamide adenine dinucleotide metabolism Hexosamine biosynthesis	Cysteine, methionine, and nicotinamide adenine dinucleotide metabolisms and hexosamine biosynthesis were aberrantly altered in PCT vs. ANT. Sphingosine was able to distinguish PCa from BHP cells for patients w/ low PSA levels. The loss of sphingosine-1-phosphate receptor 2 signaling → loss of TSG (oncogenic pathway).
Torrano et al., 2016 [232]	Peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC1α) assessment	150 = PCa 29 = control LNCaP DU145 PC3	LCHR-MS Stable isotope 13C-U6-glucose labeling	PGC1α (−) PGC1β Histone deacetylase 1	PGC1α pathway Estrogen-related receptor α (ERRα) pathway	PGC1α was a co-regulator and inhibits PCa progression and metastasis. Its deletion in murine prostate epithelium confirmed the finding. PGC1α dictates PCa oncogenic metabolic wiring, and its tumor-suppressive ability was mediated by the ERRα pathway.
Zhang et al., 2016 [233]	Angelica gigas Nakai (AGN) evaluation	5 mice per group	UHPLC-MS-MS	11 metabolites (+) including glutathione disulfide and taurine 11 metabolites (−) including lysine, tyrosine, and lactate	Methionine-cysteine metabolism Purine metabolism Citrate metabolism	Dosing w/ AGN → detectable decursinol, little decursin decursinol angelate.
Cerasuolo et al., 2015 [234]	Neuro- Endocrine transdifferen-tiation	LNCaP	H-NMR, Mathematical modeling	Creatinine + phosphor-creatinine (+) Glycine (+) Proline (+) Alanine (+) Fatty acids (+) Phospholipids (+) Glutathione (+) Glutamine (+)	Glucose oxidation Arginine and proline metabolism Glycine, serine, and threonine metabolism Glutamine and glutamate metabolism Glutathione metabolism	Hormone-deprived LNCaP cells were transdifferentiated to non-malignant neuroendocrine phenotype. Initially, LNCaP cells dwindled, neuroendocrine-type cells proliferated → later, neuroendocrine-type cells sustained LNCaP cells making them androgen-independent.
Meller et al., 2015 [235]	Metabolites analysis	106 = PCa	GC-MS LC-MS MRM	Malignant vs. non-malignant: 156 metabolites (+) 17 metabolites (−) Gleason score: 11 metabolites (+) 4 metabolites (−) ERG translocation: 53 metabolites (+) 17 metabolites (−)	Fatty acid β-oxidation Sphingolipids metabolism Polyamines metabolism Cholesterol metabolism	Fatty acid β-oxidation and sphingolipids metabolism were dysregulated in PCa relative to non-malignant tumors. TMPRSS-ERG translocated was positively correlated (causality) w/ metabolites from PCa samples. Advanced PCA tumors exhibited increased cholesterol metabolism → energy storage.

¹ The list is non-exhaustive, tabulated as of the writing of this review article. ² Total of 50 queries trimmed down to 17 integrated transcriptomic–metabolomic PCa studies.