Amino acids (AAs) are the exclusive building blocks for proteins in both the eukaryotes and the prokaryotes. It is estimated that over 500 AAs could be found in different forms of lives [1], but only 20 AAs were implicated in protein synthesis. The proteogenic AAs share a similar chemical structure with a carboxyl group and an amino group, together with a hydrogen atom, bonded to the α-carbon atom. The physical and chemical properties of different AAs are determined by the relevant side chain groups which are also covalently linked to the α-carbon atom. Owning to the chirality of the α-carbon atom, each AA has two optical isomers in theory. Interestingly, the 20 proteogenic AAs are all L-enantiomers. Previously, it was believed that AAs of D-enantiomers were absent in mammals. However, mounting evidence demonstrated that D-AAs could also be found in higher animals. For example, D-serine and D-aspartate were abundant in the central nervous system (CNS) of some vertebrates ^[2][3][2,3]. This review would pay more attention to L-AAs.

Except for protein synthesis, AAs have diverse functions under different circumstances [4]: (a) Some AAs could be converted to corresponding α-ketonic acids by aminotransferases. These α- ketonic acids could be catalyzed to glucose, fats or ketone bodies to be utilized for various purposes. The relevant AAs are called ketogenic or glycogenic AAs. (b) Some AAs behave as neurotransmitters, precursors for many hormones or other bioactive molecules. For example, glutamate is an excitatory AA in the CNS [5]. One of the enzymatic products of glutamate, the γ-aminobutyric acid, is an inhibitory neurotransmitter [6]. (c) Glycine, serine, methionine and histidine could provide one-carbon units, promoting purine, pyrimidine, and the redox regulator biosynthesis ^[7][8][7,8]. (d) Some AAs have been utilized as potential biomarkers to aid diagnosis and prognosis. For instance, elevated circulating branched-chain AAs (BCAAs) indicate a higher risk of pancreatic cancer [9]. Individuals with elevated serum concentrations of 9 AAs showed a higher incidence of Type 2 diabetes [10]. (e) AAs could behave as epigenetic modifiers to regulate cell phenotypes. Sarcosine played roles in prostate cancer diagnosis and stratification ^[11][12][11,12], and this AA also brought about increased methylated CpG island landscapes in several prostate cancer cell lines [13]. Collectively, AAs are multi-functional molecules.

Metabolites directly affect and regulate cellular phenotypes ^[14][15][14,15]. Cell transformation (tumorigenesis) is accompanied by metabolic reprogramming [16]. A half-century ago, Warburg published his profound work pointing out that tumor cells consumed a large part of glucose through glycolysis even under the circumstance of sufficient oxygen supply ^[17][18][17,18]. Several lines of evidence demonstrated that metabolic reprogramming is involved in diverse metabolic pathways when the cell transformed [19].

Glioma is one of the lethal malignancies in the CNS and accounts for one-third of the primary brain tumors [20]. The World Health Organization primarily stratified gliomas into four histopathologic grades. The higher the grade, the less favorable the prognosis [21]. The exact pathological mechanisms of gliomas are still elusive. A deep insight into the metabolic adaptation of gliomas is of benefit to exploring new therapeutic solutions.

2. Glioma AA Metabolism Adapted to Proliferation

One of the hallmarks of cancer cell metabolism is their uncontrollable proliferation abilities. This does not mean the malignant cells adopted some metabolic pathways that were unique to tumor cells. Most frequently, cancer cells change certain enzymes’ expression or activities to meet their abnormal catabolic and anabolic requirements [19]. The relevant mechanisms include the mutations of specific enzyme genes, the accumulation of specific activators/inhibitors, the altered post-translational modification activities and the over-activated/inhibited regulation signals. One-carbon unit metabolism is closely linked to many biological processes. The production of ATP, NADPH, lipids and nucleotides greatly relies on one-carbon units, especially the de novo synthesis of purine. Cell proliferation is largely determined by the availability of nucleotides [22]. Jain, M et al. profiled the extracellular consumption and release of 219 metabolites for 60 primary human cancer cell lines including U251 glioblastoma (GBM) cells ^[23][24][23,24]. They found glycine was consumed massively by rapidly proliferating cells and released by slowly proliferating ones. The uptake of extracellular glycine is mediated by the transporter of GLYT1 [25]. The intracellular biosynthesis of glycine could occur both in the cytosol and the mitochondria. Most cells default to the mitochondria for glycine synthesis [26]. Mitochondrial serine hydroxymethyltransferase 2 (SHMT2) catalyzes the reversible transformation from glycine to serine [27]. SHMT2 was frequently found to be overexpressed in GBM tissues ^[28][29][28,29]. Isotope tracing analysis indicated mitochondria contribute about two-thirds of the needed glycine to rapidly proliferating glioma cells [24]. The consumed glycine was either converted to one-carbon units to support purine synthesis or incorporated into glutathione (GSH) to clear reactive oxygen species (ROS) ^[30][31][30,31]. Glycine decarboxylase (GLDC) plays a key role in converting glycine into one-carbon units. The activity of GLDC is regulated by acetylation modification. This post-translational modification of GLDC is inhibited by the mechanistic target of rapamycin complex 1 (mTORC1). Acetylated GLDC is prone to be degraded in the proteasomes and results in impaired pyrimidine synthesis and growth inhibition of gliomas [32]. Many glioma cells exhibit highly expressed mTORC1 and GLDC ^[32][33][32,33]. γ-Glutamylcyclotransferase (GGCT), one of the key enzymes promoting GSH synthesis, was demonstrated to be highly expressed in glioma cells. Suppression or depletion of GGCT compromised glioma cells but not normal cells’ proliferation ^[34][35][34,35], implying the different GSH-related vulnerabilities of the normal and the transformed cells. In addition to GGCT, SHMT2 also promoted GSH synthesis [26]. Collectively, glycine confers gliomas a proliferative advantage by providing one-carbon units and reduction substances (Figure 1). The voracious appetite for glycine of gliomas renders glycine a promising imaging tracer to aid glioma aggressiveness evaluation [36]. Additionally, one-carbon metabolism has been identified as one of the potential targets for treating GBM [37]. Many cancer cells greatly rely on glutamine to sustain proliferation ^{[38][39][40][41]}[38,39,40,41]. Most of the tumor cells depend on aerobic glycolysis for energy production. This results in limited carbon sources entering the tricarboxylic acid (TCA) cycle. The rapidly proliferating cells need many precursors originating from the TCA cycle for building block generation. To this end, glutamine is directed to the TCA cycle to replenish the cellular carbon pool in the form of glutamate [42]. Glioma cells are glutamine-addicted both for biosynthetic and energetic purposes [43]. The increased glutamine uptake is ascribed to the highly expressed glutamine transporters. There are at least four glutamine transporters in mammalian cells. Glioma cells greatly rely on the transporters of SLC1A5 and SNAT3 to take up glutamine ^[44][45][44,45]. This feature constitutes the basis of glutamine-dependent imaging technology for glioma diagnosis and tumor boundary delineating ^[46][47][48][46,47,48]. The entry of glutamine sometimes activates mTORC1 [45], facilitating glioma cell proliferation, in part through the mechanisms described above (Figure 1). Gliomas’ glutamine addiction implies that the intrinsic synthesis capacity of glutamine is beyond glioma’s metabolic need. Whereas, Tardito, S. et al. reported that not all glioma cell lines are sensitive to glutamine starvation. In some GBM cell lines, glutamine consumption and the degree of cellular glutamine dependency exhibited no correlation [49]. Neither glutamine starvation nor glutaminase inhibition conspicuously affect cell growth. Specifically, insufficient glucose-based carbon sources from the TCA cycle to glutamate-dependent glutamine synthesis inhibited proliferation [49]. Under glutamine starvation conditions, glutamine is not directed to the TCA cycle to replenish cellular carbon sources in the form of glutamate. On the contrary, together with the alanine-originated nitrogen, the TCA cycle provides glucose-based carbon sources for glutamate production (Figure 2). The glutamine synthetase (GS) catalyzes the reaction from glutamate to glutamine. Under glutamine starvation conditions, it is the GS-dependent glutamine synthesis that fuels the de novo intracellular purine biosynthesis. GS is differentially expressed by glioma cells [49]. Gliomas harbor many genetic mutations, which might differentially affect the cell’s fate under glutamine starvation conditions ^[50][51][52][50,51,52]. For example, glutamine withdrawal from the media resulted in 80% growth inhibition of the LN18 cells. The inhibition rate of the U251 cells was three times higher than that of the LN18 cells [49]. About 50% of the GBM tissues carry EGFR mutations. EGFR could promote glioma cell proliferation through glutamate dehydrogenase-1-dependent glutaminolysis [53]. The isocitrate dehydrogenase (IDH) mutations frequently occur in gliomas and confer gliomas an extra dependency on glutamine to survive or proliferate [54]. Evidently, as of proliferation, oncogenic changes delimit glioma metabolic adaptation to some extent. Excessive ROS are frequently generated under hypoxic conditions. Gliomas outgrow their oxygen supply. This hypoxic stress impairs intracellular redox homeostasis. Under this circumstance, increased GSH synthesis is crucial for the cells to survive [55]. Glutamine deprivation not only triggers endogenous glutamine synthesis but also brought about redox stress ^[56][57][56,57]. Glioma cells cultured under glutamine starvation conditions showed upregulated expression of SHMT2 and methylenetetrahydrofolate dehydrogenase 2 (MTHFD2) [58]. Both of the enzymes participate in mitochondrial one-carbon unit metabolism. Generally, MTHFD2 expression responds to a more oxidative mitochondrial redox potential [59]. Knockdown of MTHFD2 in U87 and T98 GBM cells under glutamine-deprived conditions led to elevated cytotoxicity due to abnormal oxidation status and excessive ROS generation. Under this circumstance, serine-dependent one-carbon unit metabolism provided reducing power by promoting NADPH and GSH synthesis (Figure 2) [60]. This could be corroborated by the findings that the central areas of the resected GBM tissues (prone to suffering from hypoxia) exhibit extremely high serine and glycine content compared to the adjacent normal tissues [58]. Thus, the increased serine and glycine in the center areas of the GBM tissues indicates an active response to the requirement of both nucleoside synthesis and reducing power production (Figure 2). The metabolic effects of AAs could work synergistically. Cystine depletion could induce ferroptosis [61]. Methionine deprivation resulted in cell cycle arrest [62]. When the two AAs were depleted simultaneously, glioma cells suffered from additionally increased ROS and decreased GSH levels [63]. The double deprivation also resulted in autophagy. This synergistical effect opens up a new window to seek therapeutic solutions by manipulating multiple metabolic pathways simultaneously.

3. Glioma Migration, Invasion, and AA Metabolism

Although the metastasis of glioma is seldom found in patients, the infiltrative growth feature of gliomas is very harmful. Thus, understanding the mechanisms of glioma invasion and migration would benefit patient care. System Xc⁻ is an Na⁺-independent antiporter. It mediates the exchanging of extracellular cystine and intracellular glutamate [64]. Most gliomas highly express system Xc⁻. Glutamate excreted by gliomas is cytotoxic to the adjacent normal brain cells. Through this, gliomas create extra space to expand and invade. Lyons S. et al. reported that glutamate could activate Ca₂⁺-permeable α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid receptors (AMPA-Rs) in the manner of autocrine and/or paracrine signaling. Activation of AMPA-Rs resulted in intracellular Ca₂⁺ oscillation, one of the essential signals to trigger glioma invasion and migration (Figure 3). Most of the glioma tissues highly express AMPA-Rs. Agonists of AMPA-Rs facilitated glioma invasion and migration [65]. Additionally, glutamate showed a high affinity to N-methyl-D-aspartate receptors. Crosstalk between AMPA-Rs and N-methyl-D-aspartate receptors has been demonstrated to synergistically promote glioma invasion, especially in a glutamate-rich microenvironment. BCAA transaminase 1 (BCAT1) initiates the breakdown of BCAAs. Some catabolic products of BCAAs could be used as the carbon skeleton for glutamate synthesis. Inhibition of BCAT1 resulted in a decreased efflux of glutamate and brought about the impaired invasiveness of glioma cells [66]. In conclusion, except for its cytotoxic effects, glutamate could act as a hormone to stimulate invasion and migration (Figure 3) [67]. Many tumors show arginine dependence, and some tumors are deficient in arginine synthesis ^[68][69][68,69]. Arginine deprivation resulted in an elongated cell shape and the absence of intracellular lamellipodium, resulting in cell motility, adhesion, and invasion suppression [70]. Arginylation is crucial for actin assembly. Arginine deprivation in GBM resulted in decreased N-terminal arginylation of β-actin and brought about impaired cell motility, adhesion and invasiveness abilities [70]. Sulforaphane–cysteine, a metabolite from sulforaphane, is abundant in broccoli. It could inhibit glioma cell migration and invasion by regulating mitophagy and the expression of invasion-related molecules [71]. Factors contributing to glioma proliferation often benefit migration and invasion. For example, glutamine depletion suppressed glioma growth, invasion, and migration [72]. N-acetylcysteine is an intermediate metabolite of GSH synthesis. It could decrease the expression of Notch 2 and the relevant downstream genes, resulting in proliferation, migration, and invasion suppression. Potentially, N-acetylcysteine might induce apoptosis [73]. Thus, any intervention affecting glioma proliferation might affect invasion and migration, and vice versa.