Metabolic circuits in neural stem cells

Abstract

Metabolic activity indicative of cellular demand is emerging as a key player in cell fate decision. Numerous studies have demonstrated that diverse metabolic pathways have a critical role in the control of the proliferation, differentiation and quiescence of stem cells. The identification of neural stem/progenitor cells (NSPCs) and the characterization of their development and fate decision process have provided insight into the regenerative potential of the adult brain. As a result, the potential of NSPCs in cell replacement therapies for neurological diseases is rapidly growing. The aim of this review is to discuss the recent findings on the crosstalk among key regulators of NSPC development and the metabolic regulation crucial for the function and cell fate decisions of NSPCs. Fundamental understanding of the metabolic circuits in NSPCs may help to provide novel approaches for reactivating neurogenesis to treat degenerative brain conditions and cognitive decline.

Introduction

Increasing evidence suggests that there is a functional coupling between energy metabolism, homeostatic regulation of adult stem cells, and tissue functions [1–3]. Stem cells fine-tune the balance between self-renewal and differentiation, thereby sustaining tissue homeostasis throughout adult life. Owing to their unique microenvironment (a hypoxic niche) and their distinct cell cycle characteristics (often quiescent in adults), stem cells have a distinctive metabolic gene expression pattern and thus may rely on various metabolic pathways. Indeed, a recent metabolomics study has shown that proliferating neural stem/progenitor cells (NSPCs) in the brain had a metabolic state that was distinct from that of proliferating Schwann cells and differentiated neural cells [4]. Impaired NSPC homeostasis and/or reduced production of newborn neurons are associated with depression, dementia, and neurodegenerative disorders, such as Alzheimer’s disease. Given the therapeutic potential of endogenous NSPCs, a fundamental understanding of how metabolic changes affect cellular function and fate is necessary. To this end, this review will summarize the metabolic circuits and the key factors involved in adult NSPC homeostasis.

For more than 100 years, it was generally accepted that new neurons are not generated in adult brains. Since the late 1950s, this concept has been refuted by observations of cell division occurring in adult mouse/rat brains through in vivo labeling of newly generated cells with tritiated thymidine or synthetic thymidine analogs [5–8]. Indeed, functional neurons are continuously generated throughout life in two specific ‘neurogenic’ regions within the intact adult brain: the subgranular zone (SGZ) of the hippocampal dentate gyrus for dentate granule cells, and the subventricular zone (SVZ) of the lateral ventricle wall in the forebrain mainly for interneurons in the olfactory bulb [9–11]. These newborn neurons contribute to the structural and functional integrity of the adult forebrain [12]. A recent study showed that 700 neurons are added daily in the dentate gyrus of adult human brains [13]. Given that the cycling neuronal population is approximated to constitute 35 % of all hippocampal neurons, adult neurogenesis is a dynamic process. These adult-born neurons functionally integrate into the synaptic circuitry of the brain and perform specific functions, including olfaction, learning and memory, pattern separation, and mood control [14].

During adult neurogenesis, NSPCs go through defined stages (Fig. 1a). The criteria for developmental milestones in adult hippocampal neurogenesis include morphologies, marker expression profiles, proliferation capacities, and electrophysiological properties: (1) ‘type-1 cells’, which are the radial-glia-like stem cells, are rarely dividing predominant precursors. These cells express GFAP and nestin and give rise to type-2 cells by asymmetric division. (2) ‘Type-2 cells’, which have an irregularly shaped dense nucleus, are transiently amplifying lineage-determined progenitor cells. These cells still express nestin, but are negative for GFAP. (3) ‘Type-3 cells’, which express the immature neuronal marker doublecortin (DCX) but are nestin-negative, are migratory neuroblasts. These cells have a round-shaped nucleus and are preparing to exit the cell cycle [15, 16]. Similar to the hippocampal SGZ, three different cellular architectures exist in the SVZ. Slowly dividing GFAP-positive type B stem cells give rise to type C transit-amplifying cells, which proliferate rapidly. Subsequently, these type C cells produce type A neuroblast cells, which migrate in the rostral migratory stream towards the olfactory bulbs from the SVZ of the lateral ventricle. The majority of the cells in the adult SVZ are migrating neuroblasts that continue to proliferate [17, 18].

Fig. 1.

Schematic model of the adult NSPC niche and intracellular mechanism in NSPCs under hypoxic environment. a SVZ type B cells in this region are neural stem cells, which give rise to type C transit-amplifying cells. In turn, type C cells generate migrating neuroblasts (type A cells) destined for the olfactory bulb. Ependymal cells line the walls of the lateral ventricle. SGZ type 1 radial astrocytes give rise to type 2 transit-amplifying cells than type 3 migrating neuroblasts. Although direct measurements from oxygen tension in the SVZ have never been performed, NSCs are considered to reside in hypoxic areas of the brain. b Under the hypoxic environment, the role of HIF-1 is important. HIF-1α protein is maintained as low level under normoxia, as a result of prolyl hydroxylation-mediated degradation. However, low oxygen tension inhibits this degradation and stabilized HIF-1α translocates to the nucleus and activates gene transcription after heterodimerization with HIF-1β. Upon activation of the Notch receptor in NSCs, the Notch intracellular domain (NICD) is accumulated in the nucleus and HIF-1α potentiates Notch-dependent activation of target genes through interaction with the NICD. HIF-1α increases activation of TCF/LEF and enhances b-catenin dependent Wnt signaling. High oxygen interferes with BMP signaling. Many other signals are translated as intracellular signaling and accompanying gene expression and metabolic adaptation

Despite notable progress toward understanding NSPC biology in recent decades, various properties of NSCs remain to be identified. Whereas the tri-potency of NSCs, that is, the capacity of NSCs to generate neurons, astrocytes, and oligodendrocytes, has been well defined in culture after isolation, it remains to be elucidated whether multiple neuronal cell types can be generated by a single adult NSC. It remains controversial whether other adult brain regions, excluding the SVZ and the SGZ, have NSCs, and if so, if these NSCs can give rise to functional neurons that are ready to integrate into pre-existing neural circuits. Indeed, a recent study demonstrated the generation of interneurons within the striatum. Surprisingly, there is a robust neurogenesis and DCX-positive neuroblasts in the striatum of adult humans. Interestingly, striatal neurogenesis was less detected or not detected in patients with Huntington’s disease [19]. Because the role for olfaction is diminished in humans, adult striatal neurogenesis may provide insight on the significance of the NSPCs in the lateral ventricle wall of the brain. Additionally, the environmental cues that control region-specific neurogenesis and how those stimuli regulate the fate of NSPCs remain to be elucidated.

Metabolic adaptation to the microenvironment

The fate of stem cells is controlled by extracellular cues and intrinsic factors. Adult stem cells are hierarchically organized and maintained in a special microenvironment termed the “niche” [20]. NSCs receive signals from neighboring blood vessels, the extracellular matrix proteins found in neurogenic niches, the cerebrospinal fluid, and other brain regions (Fig. 1a). This unique microenvironment and the distinctive morphology of NSCs allow these cells to receive and integrate signals to control their cell cycle.

Response to hypoxic environment

The mammalian brain is a highly oxidative organ, and it accounts for approximately 20 % of bodily oxygen consumption. However, the partial pressure of oxygen in the brain, which ranges between 0.55 and 8 %, is lower than that of ambient air [21]. It has been proposed that stem cell niches, including neurogenic ones, are hypoxic, and stem cells have a selective survival advantage by maintaining an undifferentiated state and low oxidative stress [22]. Furthermore, NSPCs implanted into hypoxic areas of the brain maintain the viability of the surrounding neural cells through HIF-1α-mediated secretion of vascular endothelial growth factor (VEGF), which demonstrates the specialized adaptability of these cells under conditions of hypoxic insult [23]. Interestingly, multipotent NSCs and oligodendrocyte progenitors (OPCs) are more susceptible to apoptosis at 20 % oxygen than are committed neuronal progenitors, and reduced oxygen promotes multipotency [24]. In line with this finding, an in vitro culture with reduced levels of oxygen enhances survival, alters proliferation, and preconditions NSPCs toward neuronal differentiation, while maintaining them as undifferentiated NSPCs [25–27]. Notably, the dependence on hypoxia inducible factor-1alpha (HIF-1α) expression for the maintenance of the metabolic phenotype was tested in NSPCs. The deletion of exon 1 of the Hif1a in NSPCs enhanced oxidative phosphorylation during glycolytic inhibition without changing other parameters of metabolism, including their ability to survive prolonged hypoxia. This study indicated that NSPCs are intrinsically dependent on glycolysis for their survival and that HIF-1α does not mediate resistance to hypoxia [28].

Several studies have attempted to explain the regulatory links between oxygen concentration and the functions of stem cells. For example, the importance of HIF-1α-Wnt/β-catenin signaling in adult hippocampal neurogenesis in the presence of a hypoxic niche has been demonstrated [29]. Interestingly, this signaling axis does not operate in differentiated cells. Consistent with these findings, the downstream transcriptional target of Wnt signaling, for example, matrix metalloproteinase 9 (MMP9), contributes to the increased NSPC proliferation and migration [30].

To maintain the stemness, embryonic and adult stem cell populations have unique intracellular signaling and associated gene expression signatures. A lowered oxygen environment alters the signaling pathways in stem cells (Fig. 1b). The Notch pathway is conserved to regulate the stem or progenitor cell fates in most multicellular organisms [31], and hematopoietic and neuronal cell differentiation is inhibited by the Notch pathway [32]. Additionally, active Notch signaling regulates NSPC numbers through the inhibition of cell death [33]. Hypoxia activates the expression of Notch-responsive genes, such as HES1 and HEY2, by recruiting HIF-1α to the promoters of these genes [34]. The activated Notch intracellular domain enhances the recruitment of HIF-1α to its target promoters and derepresses HIF-1α function to modulate the responses to hypoxia [35]. In addition, Wnt signaling is another critical regulator of stem cells, which promotes the expansion and self-renewal of NSPCs [36]. Wnt-activated β-catenin enhances HIF-1α-mediated transcription, which suggests that this critical signaling pathway may regulate cell survival and adaptation to hypoxia [37]. In contrast, maintaining NSPC cultures in 20 % oxygen leads to mitotic arrest and promotes glial differentiation through repressing bone morphogenetic protein (BMP) signaling [38].These lines of evidence support the idea that oxygen tension dynamically regulates the developmental signaling necessary for cell fate decision and maturation and may account for the malfunction of the NSPCs during diseases and aging.

Reduced oxygen availability induces a distinct gene expression pattern in stem cells. For example, HIF-2α directly activates Oct4, an essential gene for sustaining the undifferentiated state of embryonic and adult stem cells [39]. These key genes are transcriptionally regulated by large regions of H3K27 methylation and small regions of H3K4 methylation [40, 41]. Currently, it is not entirely clear whether chromatin-based regulation is affected by the oxygen level.

Response to inflammation

A decade ago, studies have found that inflammation could block adult hippocampal neurogenesis. In particular, inflammatory factors, especially IL-6, are detrimental for NSPC survival and differentiation to neurons [42, 43]. In addition, inhibiting inflammatory microglia activation may reduce the death of newborn neurons [44] and/or attenuate hypothalamic–pituitary–adrenal axis activation [45]. Furthermore, a recent study highlights that blockade of IL-6 could largely restore hippocampal neurogenesis in a mouse model with exaggerated inflammatory responses [46]. In line with this, IL-6 and BMP act in concert to inhibit neurogenesis and promote astrocyte differentiation [47]. Because astrocytes and neuronal precursor cells share common cellular origin, it would be possible that suppression of neurogenesis by IL-6 might be due to increased astrocyte differentiation at the expense of neuronal progenitor cells. Further, it is suggested that the Janus Kinase and Signal Transducer and Activator of Transcription (JAK/STAT) 3 pathway initiated by the IL-6 family of cytokines is essential for astrocyte differentiation from NSCs [48]. Consistently, MEK, a key regulator of gliogenesis, modulates gp130-JAK/STAT3 cascade, a major cytokine-signaling pathway that promotes astrocyte differentiation [49]. Growing evidence establishes that inflammation may contribute to dysfunction of NSPCs, suggesting anti-inflammatory metabolic intervention may restore adult hippocampal neurogenesis and neuroplasticity.

Metabolic transition

Metabolic flexibility supports divergent cell fate through coordination with cellular signaling and genetic/epigenetic regulation [50]. This plasticity in the metabolic circuit is especially important to stem cells because they possess relatively low turnover rate and differentiate into specific cell types. Emerging evidence has shown that metabolic demands for maintaining stemness differ from those for differentiated cells. Dividing progenitor cells depend more on aerobic glycolysis, whereas differentiated progeny relies on energetically efficient oxidative phosphorylation [51, 52]. Morphologically, there is a loss of perinuclear mitochondrial arrangement accompanied by a low ATP/cell content, a high rate of oxygen consumption, and an increase in lipid droplets along with the loss of multipotency in differentiating primate stem cells [53].

Similarly, NSPCs show a higher glycolytic demand for energy production, proliferation and survival than do neurons in culture. The glycolytic rate of NSPC declines significantly during differentiation to neurons. Thus, NSPCs display upregulated lactate production, enhanced intracellular lactate dehydrogenase activity, and elevated glucose consumption [28]. In contrast to normal NSPCs, a recent study showed that glioma stem cells are less glycolytic than their differentiated progeny. Glioma stem cells have significantly higher intracellular ATP levels and lower glucose uptake and lactate production than do differentiated glioma cells, which support the notion that glioma cancer stem cells rely more on oxidative phosphorylation [54]. Together, the metabolic state of stem cells tends to be different from that of differentiated progeny and those under different disease contexts, including cancers.

Metabolic pathways in neural stem cells

Unique metabolic features in stem/progenitor cells

In the 1920s, Otto Warburg proposed that cancer cells generate most of their energy via the non-oxidative breakdown of glucose under aerobic conditions, instead of via the oxidation of pyruvate by the Krebs cycle. This aerobic glycolysis conflicts with the idea of the Pasteur effect, i.e., that the rate of glycolysis decreases significantly in the presence of oxygen in most mammalian cells. Emerging evidence suggests that the Warburg effect provides a biosynthetic substrate for cells that have a high need for growth and proliferation [55]. Consistently, the metabolic features of stem cells resemble the features of cancerous cells. Adult stem/progenitor cells are under high pressure to maintain tissue homeostasis by replacing damaged and expired cells: 700 new neurons are born in the adult hippocampus, and 200 billion red blood cells are produced daily. Highly proliferative embryonic stem cells (ESC) rely on glycolysis, regardless of the oxygen availability [56]. During the “reprogramming” from fibroblasts to induced pluripotent stem cells (iPSCs), epigenetic and metabolic changes occur [57]. Glucose utilization and lactate production are higher in iPSCs, whereas oxygen consumption is lower, than in parental fibroblasts. Additionally, iPSCs have shown increased levels of glycolytic enzymes and reduced levels of factors related to electron transport chains [58]. Notably, the glycolytic changes precede the expression of pluripotency markers [57]. This result suggests that somatic cells require the transition of bioenergetics from mitochondrial oxidation to glycolysis prior to reprogramming into the pluripotent state and that the metabolic changes may not be a simple consequence of acquired pluripotency. Importantly, multiple developmental pathways, such as the Notch, Sonic hedgehog and Wnt pathways, regulate stem cell self-renewal and cellular metabolism [59–63]. It is plausible that specific activation of developmental pathways may drive ‘stem cell-specific metabolic circuits’. Factors and pathways discussed here are summarized in Table 1.

Table 1.

Metabolic pathways in NSPCs

Pathways	Regulated by	Functions in NSPCs	References
Glycolysis	FoxO, TSC/mTOR, HIF-1, Sirtuin, MYC, p53	Regulate development, proliferation and differentiation	[76, 150–153, 158–161, 184–189, 195–199]
		Pivotal for cellular bioenergetics, biosynthesis, and redox homeostasis
Pentose phosphate pathway	FoxO, p53	Maintain reduced glutathione levels and suppress cellular oxidative stress by generating NADPH	[76, 77, 195, 196]
		Necessary for proliferation, growth, and migration
Glutamine metabolism	FoxO, TSC/mTOR, HIF-1, Sirtuin, MYC, p53	Maintain redox homeostasis	[76, 80–84, 125, 133, 166–168, 184, 185, 200, 201]
		Regulate growth and proliferation
		Necessary for epigenetic modifications
One-carbon metabolism		Regulate self-renewal and differentiation	[2, 85, 87]
Lipid metabolism	TSC/mTOR, HIF-1α, MYC	Regulate proliferation and neurogenesis	[4, 132–138, 154, 155, 184]

Glycolysis

In mammals, glucose is a pivotal source of cellular carbon; it is also an important fuel that is utilized in a conserved set of metabolic pathways (Fig. 2). Glucose is imported into cells via glucose transporters (GLUT) by facilitated diffusion. Intracellular glucose is phosphorylated by hexokinase (HK) to glucose-6-phosphate (G-6-P), and through glycolysis, it eventually yields pyruvates. In the cytosol, pyruvate is converted to lactate. Alternatively, upon transportation into the mitochondrial matrix, pyruvate is oxidated to acetyl-CoA or carboxylated to form oxaloacetate or malate. Pyruvate oxidation is catalyzed by the multienzyme complex pyruvate dehydrogenase (PDH), the activity of which is highly regulated by its products (acetyl-CoA and NADH) as well as by phosphorylation by PDH kinase (PDK). Acetyl-CoA, the common end product of glucose and fatty acid oxidation, conveys the carbon atoms within the acetyl group to the citric acid cycle to be oxidized for the production of GTP (or ATP) and reducing equivalents NADH2. Because citric acid cycle intermediates are used for biosynthetic pathways (e.g., amino and nucleic acids), they are constantly removed from the cycle and therefore need to be replenished through an anaplerosis reaction. At the mitochondrial inner membrane, the reducing equivalents (NADH2 and FADH) and succinate from the citric acid cycle transfer electrons to the electron transport chain for oxidative phosphorylation.

Fig. 2.

Molecular pathway of glycolysis in NSPCs. After glucose is taken up through glucose transporter (GLUT), it is converted to pyruvate by glycolysis. In turn, the pyruvate enters the energy-generating TCA cycle. HIF-1α contributes to glycolysis via turning on the expression of Glut and hexokinase (HK2). Also, HIF-1α increases the expression of pyruvate dehydrogenase kinase (PDK), which blocks the entry of pyruvate into the TCA cycle. MYC cooperates with HIF in activating several glycolytic genes. FoxO mediates glycolysis via activating transcription of glycolytic enzymes (Pgd, Pdk4) as well as ROS-detoxifying genes. The PI3K–AKT axis stimulates glycolysis by directly regulating glycolytic enzymes and by activating mTOR. The p53 suppresses glycolysis through TP53-induced glycolysis and apoptosis regulator (TIGAR) and increases mitochondrial metabolism via SCO2. G-6-P glucose-6-phosphate, PEP phosphoenolpyruvate, α-KG α-ketoglutarate, OAA Oxaloacetate, PPP Pentose Phosphate Pathway. The dashed lines indicate loss of p53 function

The differentiation of ES cells to NSPCs accompanies a relative increase in glycolysis with decreased mitochondrial oxidative phosphorylation, which suggests the importance of glycolytic regulation in NSPCs [64]. Choice of glycolytic metabolism in proliferating and neurogenic NSPCs is likely to be essential to provide building blocks necessary for cell division and growth. Glycolysis provides precursors for amino acids and nucleic acid synthesis. Hypoxic niche for NSPCs also promotes anerobic glycolysis over oxidative phosphorylation to produce ATP and to minimize oxidative stress. Although glycolysis generates only two ATP per glucose, this process is rapid enough to produce a larger fraction of ATP than mitochondrial oxidative phosphorylation [65]. Concordantly, adult NSPCs express mRNAs and proteins for glycolytic metabolism such as GLUT1 and GLUT3 [66]. Interestingly, stress conditions, such as hypoxia, regulate the densities of these glucose transporters in NSPCs. Developmental cues regulate the glycolysis necessary for neurogenesis. A key mitogen, Sonic Hedgehog, induces the expression of HK2 in cerebellar granule neuron progenitors [67]. Following glucose import, HK2-mediated glycolysis is essential to cerebellar neurogenesis.

Adult neurogenesis is impaired in obese and diabetic animal models in vivo and also by a diabetic milieu in vitro [68, 69]. As excessive glucose inhibits the proliferation and differentiation of NSPCs and as GLUT1 is required for cell growth during development, it has been hypothesized that a high level of glucose downregulates GLUT1 expression in the late stage of development and decreases glucose uptake in NSPCs. This may lead to abnormal proliferation of NSPCs and thereby results in neural tube anomalies in the embryo in pregnant diabetic animals [70]. Clearly, functional glycolysis is critical for functional NSPCs and normal development.

Pentose phosphate pathway

The glycolytic intermediate G-6-P is utilized for glycogen synthesis or the pentose phosphate pathway (PPP). The PPP in the cytosol of all cells is divided into two branches: the oxidative PPP and the non-oxidative PPP [71]. The oxidative PPP utilizes G-6-P as a substrate through the action of glucose-6-phosphate dehydrogenase (G6PD). By doing so, the primary function of the oxidative PPP is to form reduced nicotinamide adenine dinucleotide phosphate (NADPH), which is important for reductive biosynthesis and for the suppression of cellular oxidative stress by maintaining reduced glutathione (GSH) levels. In the non-oxidative PPP, the production of ribose-5-phosphate and/or xylulose-5-phosphate is important for nucleic acid and nucleotide synthesis or functions as a signaling molecule that regulates transcription [72]. In addition to the PPP, a small amount of G-6-P can enter the hexosamine biosynthetic pathway (HBP), which leads to the formation of UDP–N-acetylglucosamine (GlcNAc), a monosaccharide donor for the O-GlcNAcylation of various proteins. Because the HBP requires not only glucose but also acetyl-CoA and glutamine, it may serve as a metabolic sensor that links the metabolic status to several cellular processes [73].

An in vitro study demonstrated that NSPCs are more sensitive to pharmacological inhibition of the PPP by 6-aminonicotinamide than are differentiated neurons. This result suggests the active reliance of NSPCs on this metabolic pathway [28]. Importantly, studies in mammalian cells and yeast have shown that cellular oxidative stress redirects glucose into the PPP by a feedback regulatory loop that quells the rise in reactive oxygen species (ROS) by generating NADPH [74, 75]. Recently, we demonstrated that FoxO3 deficiency impairs the oxidative arm of the PPP by decreasing phosphogluconate dehydrogenase expression even in the face of increasing ROS [76]. This metabolic alteration may exacerbate oxidative stress in FoxO3-deficient NSPCs. Transketolase, an enzyme that mediates the non-oxidative arm of the PPP, is necessary for the proliferation, growth, and migratory abilities of cultured hippocampal NSPCs. Its knockdown or thiamine deficiency-induced inhibition of transketolase and dysfunctional PPP contributes to impaired hippocampal neurogenesis [77]. Together, these lines of evidence indicate that the importance of the PPP in the control of NSPC function is beginning to emerge.

Glutamine metabolism

Glutamine is the most abundant amino acid in plasma (~0.57 mmol/L) and intracellular muscle fluid (~19.45 mmol/L) [78]. It is an important energy source that also supplies the nitrogen for nucleotide synthesis and maintains non-essential amino acid pools. Glutamine metabolism provides a carbon source that facilitates the cell’s ability to use glucose-derived carbon and TCA cycle intermediates as biosynthetic precursors [79]. It is converted to glutamate in reactions catalyzed by a glutaminase (GLS) in which it either donates the amide nitrogen to biosynthetic pathways or is released as ammonia. Glutamate is the most abundant amino acid inside a cell that is generated by a mitochondria-associated GLS (Fig. 3). It goes through a second deamination by glutamate dehydrogenase to yield α-ketoglutarate (α-KG), a TCA cycle intermediate and substrate for dioxygenases that modifies proteins and DNA [80]. In addition, it serves as a precursor of GSH, the major endogenous antioxidant, through direct participation in the neutralization of free radicals and ROS. Glutamine, through the aspartate–malate shuttle, increases the NADPH/NADP (+) ratio, which can suppress oxidative stress [81]. Therefore, glutamine metabolism can contribute to the control of redox balance through GSH biosynthesis and/or NADPH production [82]. Furthermore, glutamine flux regulates mTOR activity, translation and autophagy to coordinate cell growth and proliferation [83]. Cellular uptake of glutamine and its subsequent efflux through SLC7A5/SLC3A2, a bidirectional transporter of glutamine-leucine and other essential amino acids, is the rate-limiting step that activates mTOR.

Fig. 3.

Molecular pathway of glutaminolysis in NSPCs. After glutamine is taken up into the cells, it is converted to glutamate by the action of glutaminase (GLS). Glutamine metabolism is important for redox balance control through GSH biosynthesis. Cellular uptake and export of glutamine is the rate-limiting step of mTOR activation. mTORC1 regulates glutamine anaplerosis. mTORC1 inhibits CREB-mediated Sirt4 transcription via proteasomal degradation of CREB2. Mitochondria-localized SIRT4 inhibits the activity of glutamate dehydrogenase (GDH, GLUD). Several transcription factors (FoxO, MYC, CREB, p53) regulate glutaminolysis through target gene activation

Adult NSPCs express GLS and incorporate glutamine into the cellular GSH pool, which is necessary for maintaining redox homeostasis [76]. Interestingly, the expression of GLS was similar in wild-type NSPCs and FoxO-null NSPCs, but the activity of GLS was altered in FoxO-null NSPCs. MYC is the most well-characterized regulator of glutaminolysis [84]. Nevertheless, how glutamine metabolism functionally controls NSPCs given MYC as a critical component of the maintenance of stem cell metabolism (see below) needs to be further characterized. Together, recent studies have reinforced the importance of glutamine metabolism in stem cell redox homeostasis, proliferation, protein and nucleic acid metabolism, and epigenetic modifications.

One-carbon metabolism

One-carbon transfers mediated by folate play essential roles in major cellular processes, including nucleic acid biosynthesis, amino acid metabolism, and vitamin and methyl group biogenesis [85, 86]. This transfer is composed of the cytosolic/mitochondrial folate cycles and the methionine cycle. The folate cycle is involved in purine and pyrimidine synthesis via the metabolism of tetrahydrofolate (THF). Folates exist in either oxidized or reduced forms, and each serves as a donor of a single carbon in different cellular reactions. The intermediate form methylene–THF is used for the conversion of dUMP (deoxyuridylate) into dTMP (thymidylate). The oxidized form of formyl–THF is utilized for the de novo synthesis of purines. Carbon from methyl–THF (reduced form) is used to methylate homocysteine to generate methionine. This reaction is catalyzed by methionine synthase (MTR) and its cofactor vitamin B12. This methionine cycle converts methionine to S-adenosyl methionine (SAM), which is a methyl donor for numerous reactions. SAM becomes S-adenosyl homocysteine (SAH), which is then converted to homocysteine.

Previous studies have demonstrated the critical role of folate in NSPC self-renewal and differentiation. Folate deficiency suppressed the proliferation of adult hippocampal NSPCs in vivo and of embryonic NSPCs in vitro [87]. NSPCs respond to folate with altered Notch signaling, increased cell proliferation [88, 89] and increased neuronal differentiation by enhancing the activity of DNA methyltransferases (DNMTs) [90, 91]. These reports suggest that one-carbon metabolism may play a critical role in NSPC fate as DNMT maintains progenitor function in self-renewing somatic tissue [92].

A recent study demonstrated that the essential amino acid threonine and SAM metabolism are coupled in pluripotent cells. Threonine, via the action of its dehydrogenase, provides glycine and acetyl coenzyme A, which together are necessary for synthesizing SAM and maintaining the SAM/SAH ratio. Threonine deprivation resulted in a decrease in SAM and a concomitant reduction in histone H3 lysine 4 trimethylation, which caused slowed growth and increased differentiation [93]. This study provides a possible mechanistic link between cellular metabolism and the epigenetic state. Whether the availability of threonine or other amino acid metabolism associated with one-carbon metabolism is necessary for in vivo NSPC function and neurogenesis warrants future investigation.

Lipid metabolism

Two major intracellular lipid metabolism pathways are: (1) the synthesis of lipids to function as structural components of cellular lipid bilayer membranes and as signaling mediators and (2) fatty acid oxidation to produce energy. Recent observations from Knobloch and colleagues highlight the importance of de novo lipogenesis in the neurogenesis of adult NSPCs [4]. Proliferating NSPCs showed higher expression levels and enzymatic activities of fatty acid synthase (FASN), the rate-limiting enzyme for lipogenesis, than did the differentiated progeny both in vitro and in vivo. Ablation of Fasn in mouse NSPCs impairs neurogenesis. Mechanistically, SPOT14, a context-dependent modulator of de novo lipogenesis [94, 95], inhibits lipid synthesis to suppress the proliferation of NSPCs by reducing the level of malonyl-CoA which is an essential substrate for FASN. Controlling FASN activity is important to ensure that stem cells maintain their pool against a premature depletion. In contrast, the fatty acid oxidation (FAO) pathway is a series of dehydrogenation, hydration, oxidation and thiolysis reactions that remove the 2-carbon acetyl-CoA from fatty acids, which then feed into the Krebs cycle to generate ATP. The importance of FAO is demonstrated as a major driver of the stem cell state by the maintenance of increased numbers of quiescent adult HSCs. FAO is active in LSK HSCs, whereas it is undetectable in differentiated cells. Inhibition of mitochondrial FAO by deletion of Ppard, a critical regulator of the transcriptional program underlying FAO, reduced functional HSCs, whereas pharmacological PPARδ activation improved HSC maintenance. Mechanistically, the inhibition of FAO increased symmetric cell divisions, which results in a premature depletion of HSCs [96]. Whether FAO is a functionally essential metabolic component in the maintenance of quiescent adult NSCs in vivo remains to be determined. In addition, cholesterol, one of the key lipid constituents of the plasma membrane, is important for the formation and maintenance of the CNS. Ablation of cholesterol biosynthesis through conditional knockout of squalene synthase in NSPCs reduced the size of the embryonic mouse brain [97]. A recent study reported that many genes expressed by SOX2-positive NSPCs of DG are enriched in gene ontology categories associated with the generation of energy, lipid metabolism and catabolic processes such as Bmp6, Apoe, Acacb, Igfbp7, Edf1, Irs, Lrp1, Cpt1a, Pparc1a, Acaa2, and Acsl6 [98]. This finding is in line with previous studies, which suggest the emerging importance of a lipid metabolic program in DG NSPCs.

Factors regulating metabolic circuits in neural stem cells

RTK–PI3K/PTEN–PDK1–Akt pathway

Neurogenesis is greatly dependent on extracellular cues [99, 100]. The regulation of NSPC proliferation by epidermal growth factor and fibroblast growth factor signaling has been well described [101]. Insulin-like growth factor (IGF)-1 increases the proliferation and neuronal differentiation of NSPCs [102]. Consistently, transgenic mice that overexpress IGF-1 have larger brains with greater numbers of neurons [103]. Additionally, the peripheral in vivo infusion of IGF-1 stimulates neurogenesis in the adult rat hippocampus [104]. The PI3K–Akt pathway is downstream of receptor tyrosine kinases and regulates the proliferation and cell fate of NSPCs. A recent study demonstrated that endogenous ROS enhances self-renewal and neurogenesis both in vitro and in vivo in a PI3K–Akt pathway-dependent manner [105]. The importance of this pathway in NSPCs is further supported by multiple lines of in vivo evidence. Phosphatase and tensin homolog (PTEN) functionally antagonizes the PI3K–Akt pathway. Brain-specific PTEN null mice showed a dramatically increased brain volume through increased cell proliferation, decreased cell death, and enlarged cell size. In addition, PTEN restricts self-renewal of NSPCs both in vitro and in vivo [106, 107]. Phosphoinositide-dependent kinase 1 (PDK1) is an upstream activator of Akt and contributes to the early generation of OPCs from NPCs [108]. Ablation of PDK1 in NSCs resulted in a reduction in the number of SOX10 and platelet-derived growth factor receptor α (PDGFRα)-positive OPCs within the neocortex and the striatum during embryogenesis. Furthermore, the PDK1–Akt pathway regulates, through activation of Mash1, neuronal differentiation and subtype specification during telencephalic development [109]. The reduced brain size and weight with the smaller ventricular system in PKB gamma/Akt3-deficient mice suggest that Akt is crucial for postnatal brain development [110]. Deletion of GSK-3, a downstream effector of Akt, resulted in massive hyperproliferation of NSPCs at the expense of both intermediate neural progenitors and postmitotic neurons. These effects were associated with the dysregulation of Wnt, Sonic Hedgehog, Notch and fibroblast growth factor signaling [111]. Together, these lines of evidence have shown that the RTK–PI3K/PTEN–PDK1–AKT–GSK3 axis is a critical regulator of NSPC homeostasis during brain development.

The PI3K–Akt pathway constitutes a major regulator of cellular metabolism. The activation of this pathway upregulates glucose, glutamine and lipid metabolism through downstream effectors such as FoxO, mTOR, HIF-1α, Myc, and p53, as discussed below. In addition, Akt directly regulates a number of metabolic enzymes [112], representing a master regulator of the metabolic network within a cell. For example, Akt stimulates glycolysis by increasing the expression and membrane translocation of glucose transporters and by phosphorylating key glycolytic enzymes, such as hexokinase (HK) and phosphofructokinase 2 (also known as PFKFB3) [113, 114]. Akt also activates ectonucleoside triphosphate diphosphohydrolase 5 (ENTPD5), an enzyme that supports increased protein glycosylation in the endoplasmic reticulum and indirectly increases glycolysis by creating an ATP hydrolysis cycle [115].

FoxO

As a key downstream effector of the PI3K–Akt pathway, FoxO transcription factors (FoxOs) are involved in cellular proliferation, differentiation, survival, ROS detoxification and metabolism [116–118]. Increasing evidence has indicated that FoxOs are crucial in correctly reinforcing stem cell fates. FoxO1 regulates pluripotent self-renewing circuitry by directly activating the expression of Oct4 and Sox2 in human ESCs [119]. In addition, FoxO4 is necessary for the neural differentiation of ESCs [120]. FoxO1 is also essential for spermatogonial stem cell self-maintenance and spermatogenesis [121]. In particular, FoxOs play important roles in adult stem cell homeostasis and maintenance. FoxO deficiency in HSCs results in the accumulation of ROS and the hyperproliferation of the HSC population, which leads to the reduction of the long-term repopulating capacity of HSCs [118, 122]. Consistently, the targeted ablation of FoxO isoforms has resulted in the early depletion of NSPCs, which further supports the importance of FoxO in maintaining the regenerative potential of adult brains [123]. A previous study reported that FoxO3 regulates a range of genes involved in glucose metabolism and the adaptation to hypoxia in NSPCs and that its ablation resulted in age-progressive NSPC depletion [124]. A recent endeavor to link metabolic regulation by FoxO to NSPC fate has revealed that FoxO3 regulates glucose and glutamine metabolism to suppress oxidative stress in NSPCs [76]. FoxO3-deficient NSPCs had decreased levels of mitochondria-localized HK2 through reduction of pAKT, which eventually resulted in the disruption of glucose uptake. Additionally, accumulated ROS by FoxO3 ablation lowered the activity of pyruvate kinase M2 (PKM2), a rate-limiting glycolytic enzyme, which also contributes to the glycolysis defect. The increase in ROS is due, in part, to altered glutamine metabolism. The decreased incorporation of ¹³C- isotope labeled glutamine-derived glutamate into the cellular GSH pool explains that decreased GLS activity accounts for oxidative stress in FoxO3-deficient NSPCs [76]. Together, this study emphasized the importance of glucose and glutamine metabolism in neural stem cell maintenance.

Notably, a recent study by van der Vos et al. identified glutamine synthetase (GLUL) as a transcription target of FoxOs [125]. The activation of FoxO3 and FoxO4 upregulates the expression and activity of GLUL, which, in turn, results in the elevation of glutamine levels. Furthermore, FoxO-mediated glutamine upregulation contributes to the inhibition of mTOR signaling and the induction of autophagy. Autophagy is a lysosomal degradation pathway that allows cells to recycle damaged organelles for energy and biosynthesis. Deficiencies in this pathway are associated with neurodegeneration and impair neuronal differentiation. Autophagic flux and LC3 lipidation increase, along with the increased expression of Atg7, Becn1, Ambra1 and LC3, are required for the early neuronal differentiation of embryonic NSPCs. Consistently, Atg5- and Ambra1-deficient NSPCs display defective neuronal differentiation [126]. Ablation of FIP200, a gene essential for autophagy induction, resulted in p53 activation and a progressive loss of NSPCs and impairment in neuronal differentiation specifically in the postnatal brain [127]. It remains to be determined whether FoxO-dependent glutamine metabolism contributes to NSPC differentiation through supporting autophagy.

TSC/mTOR

TSC-regulated mTOR is necessary for the maintenance and functions of NSPCs by regulating self-renewal, differentiation, and terminal maturation. Cortical embryonic NSPCs of Tsc1 ^L/L Emx1-Cre mice showed defective self-renewal due to the attenuation of PI3K–Akt activation [128]. This is due to the hyperactivation of mTORC1-S6K, which causes negative feedback regulation of the upstream PI3K–Akt. Consistent with the defects seen in other stem cell compartments, mTOR hyperactivation increased NSPC numbers, followed by premature neuronal differentiation and impaired maturation, both during embryonic and postnatal development. Consistent with these findings, the targeted deletion of Tsc1 in embryonic NSPCs has led to megalencephaly and the formation of giant cells in the cerebral cortex [129, 130]. Furthermore, the ablation of Tsc1 in postnatal SVZ NSPCs by a Nestin-CreERT2 transgene caused nodular masses outgrowing into the ventricles [131]. Together, these studies demonstrated that aberrant mTOR hyperactivation in NSPCs leads to deleterious dysregulation.

mTOR integrates growth signals and nutrient availability. The activation of mTOR leads to protein and lipid biosynthesis and cell growth [132]. Mechanistically, mTOR directly stimulates mRNA translation and ribosome biogenesis and indirectly causes metabolic changes by altering other key regulators. For example, mTOR activates the translation of HIF-1α even under normoxic conditions, thereby upregulating glucose uptake and aerobic glycolysis (see below). Furthermore, a recent study demonstrated that mTORC1 promotes glutamine anaplerosis by activating glutamate dehydrogenase (GDH). This regulation requires transcriptional repression of SIRT4, which is the mitochondria-localized SIRT that inhibits GDH. Mechanistically, mTORC1 represses SIRT4 by promoting the proteasome-mediated destabilization of cAMP-responsive element binding 2 (CREB2) (Fig. 3) [133]. Additionally, mTOR signaling regulates lipid homeostasis [134, 135]. mTOR upregulates cholesterol biosynthetic gene expression and lipogenesis, whereas it suppresses lipolysis and fatty acid oxidation [136]. In contrast to its effect in vitro, Rapamycin treatment in vivo leads to dyslipidemia and hypercholesterolemia in patients due to defects in lipid mobilization and transport [137]. Whether mTOR-regulated lipid metabolism is operative in NSPCs outside adipose and hepatic tissues remains to be determined.

HIF-1α

HIF-1α is responsible for the transcription of hypoxia-regulated genes. In most cells, HIF-1α is constitutively produced but rapidly degraded. Hydroxylation of HIF-1α by prolyl hydroxylases recruits the VHL E3 ubiquitin ligase, which leads to proteolytic destruction. To date, oxygen-regulated HIF-1α has been shown to play a critical role in regulating glucose and glutamine-derived central carbon utilization in mammalian cells [138].

Adult stem cells, including NSPCs, reside in hypoxic niches, and the development states of these cells are affected by oxygen availability. Therefore, it is conceivable that oxygen-sensitive regulator HIF signaling is closely related to their fates [139]. Indeed, studies performed in HSCs generally support the critical role of HIF-mediated function in maintaining stem cells. The disruption of HIF-1α/ARNT heterodimers through the depletion of Arnt led to failed hypoxia-mediated progenitor proliferation as well as a decreased hematopoietic progenitor population during development [140]. Furthermore, HIF-1α deletion progressively increased cell cycling and decreased long-term repopulation capacity in adult HSCs [141]. Quiescent HSCs slowly cycle, causing low-level oxidative stress. This prevents undue differentiation and premature exhaustion of HSCs [142]. Therefore, HSCs maintain intracellular hypoxia and a suitable amount of HIF-1α for their quiescence. However, the result that the overexpression of HIF-1α impaired the transplantation capacity suggests the requirement of a fine-tuned regulation at the HIF-1α level.

As mentioned earlier, Hif-1α deletion impairs hippocampal neurogenesis by suppressing NSPC proliferation, differentiation and neuronal maturation. This defect correlates with a decline in Wnt/β-catenin signaling in the SGZ, which clearly suggests the critical role of HIF-1α in NSPC biology [29]. A number of studies have demonstrated that intermittent hypoxia increased neurogenesis in the adult rat SVZ and hippocampus, thereby contributing to memory and learning [143–145]. Hypoxia increases cyclin D1 expression through activation of JNK in rat NSPCs in vitro, suggesting a novel mechanism for hypoxia-induced proliferation of NSPCs [146]. This occurs, in part, by HIF-1α-dependent proliferation of NSPCs and subsequent increases in newly generated neurons [147]. Studies using hyperbaric oxygen treatment also reported enhanced proliferation and neuronal differentiation of NSPCs [148]. These effects are presumably due to the stabilization of HIF-1α, as previously reported [149].

HIF-1α regulates multiple steps of glycolysis [150]. Extracellular glucose is imported into cells through GLUT1, the expression of which is regulated by HIF-1α [151]. Glucose is, in turn, phosphorylated by hexokinase to generate G-6-P. HK2, which is a hexokinase isoform and a key mediator of aerobic glycolysis, is a target gene of HIF-1α [152]. HIF-1α, in cooperation with c-MYC, induces pyruvate dehydrogenase kinase 1 (PDK1) and suppresses oxidative phosphorylation by blocking the entry of pyruvate into the TCA cycle [142]. As described previously, HSCs maintain stabilized HIF-1α, which leads to elevated rates of glucose consumption and lactate production and reduced levels of mitochondrial oxidative phosphorylation and oxygen consumption [153].

HIF-1α directs glutamine-derived carbons away from the TCA cycle to be used for fatty acid synthesis [138]. This isocitrate dehydrogenase 1 (IDH1)-dependent pathway of the reductive carboxylation of glutamine-derived α-KG for de novo lipogenesis is particularly critical for cellular proliferation under hypoxia [154]. A recent report forged a mechanistic link between hypoxia-suppressed glutamine oxidation and lipid synthesis [155]. The activation of HIF-1α activates the E3 ubiquitin ligase SIAH2, which induces the degradation of a subunit of the mitochondrial enzyme α-KG dehydrogenase. Decreased glutamine oxidation counter activates its reductive carboxylation to citrate and fatty acid synthesis. Together, accumulating evidence supports the idea that HIF-1α facilitates metabolic adaptation upon redox change and the mobilization of NSPCs to proliferate, differentiate, or migrate in and out of the hypoxic niche.

Sirtuin

Sirtuin is a family of nicotinamide adenine dinucleotide (NAD)-dependent deacetylases. Thus, it integrates cellular energy status and stress. SIRT1 and 2 are present both in the nucleus and the cytoplasm in different cellular contexts, whereas SIRT3, SIRT4, and SIRT5 are mitochondrial and SIRT6 and SIRT7 are constitutively nuclear [156, 157]. Recent studies have emphasized its role in cellular metabolism and energetics. Studies in knockout mice have revealed that the SIRT family regulates lifespan, genomic stability and metabolism [157].

SIRT1 activates PGC-1α, the master regulator of mitochondria biogenesis, and inhibits glycolysis [158]. Consistently, SIRT1 promotes cellular autophagy, a process that is tightly coupled with mitochondria metabolism [159]. Similarly, recent studies have demonstrated that SIRT3 and SIRT6 deacetylate and inactivate HIF-1α and thus suppress glycolytic metabolism [160, 161].

In particular, SIRT3 is the major mitochondrial deacetylase, and multiple targets have been identified and shown to be important for cellular metabolic homeostasis. For example, SIRT3 deacetylates OXPHOS complexes I, II, and III, consequently increasing mitochondrial aerobic respiration. SIRT3 knockout mice also display reduced mitochondrial oxidation and increased ROS [162]. Furthermore, it deacetylates and activates IDH2. Thus, SIRT3 also protects cells from oxidative stress by increasing the GSH/GSSG ratio [163]. Superoxide dismutase (SOD) is another substrate of SIRT3 that is activated upon caloric restriction [164]. A recent study demonstrated that the expression of SIRT3 regulates HSC self-renewal under stress or at an old age by maintaining mitochondrial metabolic homeostasis and suppressing ROS [165]. These findings suggest SIRT3 is critical for the suppression of cellular oxidative stress and necessary for mitochondria metabolism.

As described previously, SIRT also has a profound role in glutamine metabolism. SIRT4, through its ADP ribosylation of GDH, inhibits glutamine metabolism in the mitochondria [166]. In contrast, SIRT3 has been proposed to activate GDH activity through its deacetylation [167, 168]. Interestingly, SIRT3-deficient mice are metabolically unremarkable under basal conditions, suggesting that it may have cellular context-dependent activity. How different SIRTs oppositely regulate GDH and orchestrate cellular metabolism requires further investigation.

In NSPCs, SIRT1 is expressed and regulated during self-renewal and neurogenic differentiation [169]. The knockdown of SIRT1 does not affect proliferation but does enhance the neuronal differentiation of NSPCs. Another study demonstrated that SIRT1 responds to redox stress and alters the NSPC cell fate decision toward preferred differentiation into the astroglial lineage. The mechanism by which SIRT1 is activated by ROS is unclear, but it mediates deacetylation of histone H3K9 in the MASH1 promoter by SIRT1 and the subsequent downregulation of the pro-neuronal transcription factor MASH1 [170]. In contrast, a separate study reported that SIRT1 represses the Notch1-Hes1 pathway, and its transient translocation into the nucleus is necessary for the neuronal differentiation of NSPCs [171]. Notably, a recent study that exploited a conditional SIRT1 knockout driven by the Nestin-CreERT2 transgene described that the main phenotype of SIRT1-deficient NSPCs is the expansion of OPCs in the adult SVZ [172]. Together, SIRT1 clearly has an indispensable function in NSPC fate decision. How developmental cues, aging, and the level of stress influence the activity of SIRT1 as well as other SIRT family members in NSPCs and the outcome warrants further investigation.

Of note, SIRT is regulated by dietary restriction, pharmacological interventions and intracellular redox potential [173]. Its activity may be mediated through other factors such as FoxO, p53, Myc, and HIF-1α. Identifying its cellular substrates and the mechanism of gene expression changes is an active area of investigation given its role in adult stem cell homeostasis and organismal aging.

MYC

MYC is necessary for the normal development of the neural system: in the mouse, knocking out c-myc is lethal at embryonic day 10.5 with defects in neural development [174], and the reduced expression of N-myc disrupts early embryogenesis and postnatal brain growth [175]. Confirmation of MYC expression during fetal brain development [176] and in NSPCs [177] implies that MYC also contributes to the NSPC fate decision. Indeed, a fraction of the MYC-overexpressing cells retain the self-renewal capacity even under differentiation-inducing culture conditions. Conditional deletion of N-myc in nestin-expressing NSPCs decreases the whole brain size and obstructs cerebellar development due to the disrupted neuronal differentiation [178]. Deletion of either c- or N-myc or both genes in NSPCs driven by nestin–cre causes microcephaly, which accompanies a strong reduction in the cerebellar granule neural progenitors and mature granule neurons, phenocopying the genetic disorder of Feingold syndrome [179]. MYC-deficient brains exhibit the downregulation of genes involved in protein and nucleotide metabolism, mitosis, and chromatin structure as well as the upregulation of genes associated with differentiation [180]. These findings support the notion that MYC has a profound role in the cell fate decision of NSPCs by inhibiting differentiation as well as directing active cellular metabolism.

MYC controls self-renewal and differentiation in adult stem cells, but its function differs depending on the types of stem cells and the contexts. In bone marrow, the conditional deletion of c-myc was found to lead to the accumulation of HSCs mainly as a result of differentiation failure [181]. In contrast, the knockout of MYC in mammary basal cells was found to result in the impairment of self-renewal and the diminishment of the progenitor population [182]. Unexpectedly, MYC promotes the differentiation of NSPCs, rather than their proliferation [183]. The depletion of endogenous MYC in radial glial precursors inhibits differentiation, whereas the overexpression of MYC induces neurogenesis. This finding is in contrast to previous reports that demonstrated the mitogenic activity of MYC in dissociated cells in culture. In this recent study, MYC was shown to drive differentiation by inhibiting Notch signaling and by increasing neurogenic cell division, eventually resulting in a depletion of progenitor cells within the intact polarized tissue [168]. These findings indicate an unexpected role of MYC in the control of NSC stemness versus differentiation in vivo.

MYC is critical in the transcriptional and post-transcriptional regulation of stem cell maintenance. MYC-mediated metabolic regulations in stem cells remain an active area of investigation. It has been suggested that MYC controls embryonic and adult stem cell physiology, in part, through the regulation of cellular glutamine metabolism [184]. MYC stimulates mitochondrial glutaminolysis through coordinated transcriptional and post-transcriptional programs. It transcriptionally activates two main glutamine transporters, SLC1A5 and SLC38A5, which results in the reprogramming of mitochondrial metabolism to depend on glutamine catabolism. This MYC-dependent glutaminolysis stimulation reduces the entry of glucose carbon into the TCA cycle and triggers cellular addiction to glutamine as an energy source [185]. In addition, MYC enhances mitochondrial GLS expression and glutamine metabolism by suppressing microRNAs miR-23a/b [84].

MYC enhances aerobic glycolysis by directly regulating the transcription of glycolytic genes [186]. Most glycolytic genes have conserved MYC-binding E boxes (ENO1, HK2, or LDHA) or can be induced by MYC even without canonical E boxes (GAPD and TPI1) [187]. In addition to transcriptional regulation, MYC has been shown to promote aerobic glycolysis in cancer cells by promoting RNA splicing through the controlled expression of hnRNP splicing factors [188]. In glioblastoma, mTORC2 was shown to regulate glycolytic metabolism through c-MYC [189]. It remains to be determined whether suppressed glycolysis is the major cause of NSPC proliferation defects upon MYC ablation. Together, these lines of evidence indicates that MYC regulates key aspects of NSPC function, including ribosomal and mitochondrial biogenesis, glucose and glutamine metabolism, lipid synthesis, and cell cycle progression [184].

p53

The expression of p53 within NSPC-rich neurogenic regions has been reported [190]. p53 regulates NSPCs at multiple levels. Knockout mouse studies have suggested its role in cell death during postnatal hippocampal neurogenesis [191]. The loss of p53 enhanced adult SVZ NSPC self-renewal [192]. This finding is consistent with the knockout phenotype of p21, its bona fide transcriptional target [193]. p53 deletion in NSPCs and the telencephalon resulted in an elevation in ROS and PI3K–Akt signaling, which was linked to enhanced neurogenesis. By contrast, the downregulation of ROS levels and PI3K signaling in p53 ^−/− NSPCs ameliorated the defect [194]. Simultaneously, the reported effect of an in vitro culture condition with 20 % oxygen on the decreased proliferation and enhanced glial differentiation of NSPCs was ascribed to p53 activation [24, 38].

In agreement with in vivo evidence, the biochemical characterization of p53 ^−/− cells suggested its role in maintaining cellular redox potential. First, p53-induced glycolysis and the apoptosis regulator (TIGAR), a p53-inducible fructose 2,6-bisphosphatase, antagonize glycolysis. TIGAR redirects glucose to the PPP and increases the production of NADPH required for maintaining reduced GSH, which leads to the subsequent reduction of cellular ROS levels [195, 196]. Consistently, p53 ^−/− mice display enhanced glycolysis and an increase in ROS [197, 198]. Furthermore, p53 regulates the balance between mitochondrial respiration and glycolytic metabolism. This regulation is accomplished by activating synthesis of cytochrome c oxidase 2 (SCO2), which is critical for the assembly of the cytochrome c oxidase (COX) complex [199].

p53 transcriptionally activates mitochondrial glutaminase 2 (GLS2) as part of DNA damage responses [200, 201]. GLS2 increases the mitochondrial conversion of glutamine to glutamate, which contributes to the production of GSH and energy. In addition to these findings, p53 engages in crosstalk with the PI3K–Akt pathway through multiple avenues and indirectly contributes to cellular metabolism. For example, p53 has been shown to upregulate PTEN and to suppress PI3K–Akt signaling, thereby opposing the metabolic activity driven by this major regulatory pathway [202].

Metabolic control of epigenetic alteration in NSPCs

Recently, cellular metabolism and the epigenetic control of gene expression, especially DNA and/or histone methylation, have been reported to have an important role in the differentiation process. Isocitrate dehydrogenase 1 (IDH1) and its mitochondrial homolog IDH2 are notable examples. Mutations in IDH1/2 are most frequent in acute myeloid leukemia (AML) and the gliomas in different human cancers [203, 204]. The neomorphic IDH1/2 mutations in tumors catalyze a NADPH-dependent reduction of α-KG to R-2-hydroxyglutarate (2-HG) [205]. This metabolic alteration results in DNA hypermethylation through the inhibition of TET2, an α-KG-dependent ten–eleven translocation (TET) family of DNA hydroxylases. A recent study demonstrated a proleukemogenic activity of 2-HG that impairs hematopoietic differentiation by inhibiting TET2 [206, 207]. Additionally, 2-HG competitively inhibits the α-KG-dependent Jmjc domain that contains histone demethylases, which suggests an explanation for the altered abundance of histone methylation marks in mutant IDH-expressing cells [208, 209]. Consistent with these proposed mechanisms, IDH1/2 mutations also alter cell fate determination in NSPCs. The enforced expression of mutant IDH1(R132H) interferes with neural differentiation, and this differentiation blockade of non-transformed cells is based on the enhancement of repressive H3K9 methylation and CpG-island hypermethylation [210]. Consistent with previous reports that suggest a role for IDH1/2 mutants in epigenetic reprogramming [210, 211], a recent study reported that IDH mutant-specific small-molecule inhibitors induce the expression of genes and cell markers associated with glial-specific differentiation in glioma cells and the reduction of repressive chromatin marks (H2K9 and H3K27 trimethylation) at the promoters of induced genes [212].

Conclusion

As exemplified in the aforementioned examples of metabolic regulation of NSPC fate, clinical testing of targeted metabolic therapies to boost neurogenesis is well justified. Many pharmacological interventions for metabolic syndromes and abnormalities are currently available. The identification and characterization of new targets is imperative for developing novel therapeutic strategies that exploit the regenerative potential of resident NSPCs against neurodegenerative diseases. To this end, additional studies are needed to demonstrate the functional significance of epigenetic modifications and specific gene expression patterns affected by cellular metabolism.

Additional evidence supports the possibility that grafted NSPCs may replace damaged or injured cells and improve motor function in neurodegenerative conditions [213]. In particular, studies are now seeking to tailor cell replacement therapy to each patient. Currently, the ex vivo manipulation and expansion of NSPCs is a popular approach to understanding this intriguing cell type. A better understanding and design of the most suitable microenvironment-mimicking culture conditions will ensure better progress toward generating neural cells for the various therapies that are under development.