Two-way communication between the metabolic and cell cycle machineries: the molecular basis

Abstract

The relationship between cellular metabolism and the cell cycle machinery is by no means unidirectional. The ability of a cell to enter the cell cycle critically depends on the availability of metabolites. Conversely, the cell cycle machinery commits to regulating metabolic networks in order to support cell survival and proliferation. In this review, we will give an account of how the cell cycle machinery and metabolism are interconnected. Acquiring information on how communication takes place among metabolic signaling networks and the cell cycle controllers is crucial to increase our understanding of the deregulation thereof in disease, including cancer.

Resting cells require a basal level of catabolic metabolism to ensure energy homeostasis. Cells that commit to entering the cell cycle, however, differ greatly from resting cells in terms of their metabolic profile, as they will eventually have to double their cell content, that is, their DNA, membranes, organelles and other biomass. To support the energy-consuming processes needed for this program, cells increase the uptake of glucose and glutamine and shut down oxidative metabolism. In this way, glucose and glutamine-derived metabolic intermediates can be used for the biosynthesis of macromolecules required for the cell division. Highly proliferating cells, including cancer cells but also activated lymphocytes, thymocytes and embryonic cells, preferentially use glycolysis even in the presence of oxygen.^1-8 This phenomenon is called aerobic glycolysis or “the Warburg effect”.⁹

In unicellular organisms, cell cycle progression is dependent on the availability of nutrients, which directly couples available resources to the generation of progeny. For example, stationary-phase yeast switches to a mitotic phenotype when exposed to glucose, but becomes quiescent or sporulates when no other nutrients are provided.¹⁰ Under nutrient-steady growth conditions, cycling yeast cells display fluctuations in oxygen consumption, alternating between glycolysis and respiration. Their cell division is solely limited to the glycolytic phase, with DNA replication taking place only during that period.¹¹ Interestingly, many genes identified in classic screens for factors regulating the cell cycle in yeast, were later shown to have a function in metabolic regulation, too.^12-17 Also, transcriptome studies demonstrated that yeast genes involved in glycolysis respiration, lipids and amino acid synthesis are expressed as a function of the cell cycle.^18,19 Taken together, these observations show that in unicellular organisms, intimate connections between cell cycle and metabolism must exist.

In contrast to single-cell eukaryotes, cells of multicellular organisms usually have an unlimited access to nutrients. However, they are not cell-autonomous for nutrient uptake but instead depend on proliferation-regulating pathways. Mitogen-mediated activation of signaling routes triggers nutrient uptake and represents the rate-limiting cue for cell cycle entry.²⁰ As a consequence, growth factor-stimulated cells initiate cell division in a fashion comparable to that of yeast exposed to a nutrient-rich medium.^21,22 Accordingly, in the absence of mitogens, even in a nutrient-rich environment, cells will not enter the cell cycle.²³ On the other hand, even in the presence of promitogenic cues, glucose deprivation will keep cells from proliferating, which is a widely used method for synchronizing mammalian cells.²⁴ The fact that signaling pathways coordinating cell cycle progression control, and are controlled by, changes in cellular metabolism^25,26 shows that, also in multicellular organisms, there must be a crosstalk between these pathways, cell cycle and metabolism. Yet, the molecular basis that connects nutrient availability, biosynthetic intermediates and energetic balance to the core cell cycle machinery remains incompletely understood. Here, we will give an overview of how the cell cycle machinery and metabolism are interconnected.

Cell Cycle Regulation of Metabolism

Evidence is emerging in support of the coordinated temporal regulation of metabolism directly by the cell cycle modulators. A first indication for this came from the observation that in yeast, metabolites of nucleotide, protein and lipid synthesis are cyclically fluctuating, as a function of cell cycle progression.²⁷ Indeed, it has been shown subsequently that the glycolysis-promoting enzyme 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3 (PFKFB3) is subjected to cell cycle dependent temporal regulation by members of the ubiquitin proteasome system (UPS; Figure 1, upper panel).^28,29 Since then, a number of mechanisms have been revealed that couple the cellular metabolic state to the cell cycle (Fig. 2).

Figure 1.

Protein activities and metabolic events during the cell cycle. A schematic representation of the temporal regulation of metabolic factors (upper panel) and the cell cycle machinery (lower panel). The represented protein levels are not relative, but rather indicate their relative timing of expression.

Figure 2.

Bidirectional regulation of the cell cycle machinery and metabolic enzymes. Interactions between the cell cycle machinery and metabolic enzymes indicated by arrows may be either direct or indirect. Color codes represent proteins belonging to the cell cycle machinery or to metabolism.

Most somatic cells are differentiated and quiescent, that is, they reside in the G0 phase of the cell cycle. Following mitogenic stimulation, cells typically re-enter the cell cycle and proceed through the G1 phase, in which the stage is set for DNA replication. Upon passage through the G1/S restriction point, cells enter S phase in which they double their DNA content, move on into the G2 phase and the final mitotic (M) phase, in which cellular contents are divided over 2 daughter cells (Fig. 2). Key proteins for the tight regulation of the cell cycle are cyclin-dependent kinases (CDKs), which associate with one of different cyclins across the cell cycle to ensure accurate cell cycle progression.^30-33 The kinase activity of cyclin-CDK complexes is tightly regulated by a plethora of CDK inhibitors (CKIs), which stop cell cycle progression in unfavorable circumstances.³⁴

D-type cyclins

The role of D-type cyclins in metabolism was first demonstrated in cyclin D-deficient mice that display marked metabolic phenotypes. Cyclin D2-deficient mice show a diabetic phenotype due to impaired pancreatic β-cell expansion and function, which is further enhanced by cyclin D1 co-depletion.³⁵ Cyclin D3-deficient mice display reduced adipocyte size and increased sensitivity to insulin, which is a consequence of the inactivation of peroxisome proliferator-activated receptor (PPAR)γ, the master regulator of adipogenesis (Fig. 2).³⁶ Accordingly, depletion of cyclin D3 diminishes PPARγ activity and adipogenesis, while cyclin D3 overexpression has the opposing effect. Cyclin D1 is highly expressed in breast cancer cells; among other functions, it decreases the abundance and activity of the glycolytic enzyme hexokinase 2 (HK2; Fig. 2).³⁷ Correspondingly, cyclin D1 depletion in either normal or oncogenic breast cells leads to an increase in glycolytic enzyme pyruvate kinase (PK) as well as the lipogenic enzymes acetyl-CoA carboxylase (ACC) and fatty acid synthase levels (FASN; Fig. 2).³⁷ In hepatocytes, cyclin D1 inhibits the glucose-mediated induction of central lipogenic genes via repression of the carbohydrate response element binding protein (ChREBP) and hepatocyte nuclear factor 4α (HNF4α), which are important regulators of glucose sensing and lipid metabolism (Fig. 2).³⁸ Hence, cyclin D1 inhibits both glycolysis and lipogenesis. Moreover, it hinders mitochondrial biogenesis and functions through inhibition of the nuclear respiratory factor 1 (NRF1), which regulates nuclear-encoded mitochondrial genes, and the mitochondrial voltage-dependent anion channel (VDAC), respectively (Fig. 2).^39,40 Finally, unbiased mass-spectrometry analysis of proteins interacting with cyclin D1 using cyclin D1 knockin mice revealed interactions between cyclin D1 and numerous metabolic proteins, such as lipogenic enzymes FASN and ACC, as well as the mitochondrial electron transport chain components cytochrome c oxidase (COX) and ATP synthase (Fig. 2).⁴¹ These observations together demonstrate that cyclin D1, by hindering glycolysis, lipogenesis as well as mitochondrial activity, downregulates metabolic activity through several routes.

Restraining the conversion of glucose to lipids for storage by D-type cyclins might allow cells to use glucose-derived metabolites for doubling their cell content for cell division to take place. Alternatively, cyclin D-mediated inhibition of metabolism might provide a negative feedback loop to ensure unidirectionality of cell cycle progression.⁴² Cyclin D activity accumulates in G1, as a function of growth factor signaling, hence when metabolic activity is abundant (Fig. 1).⁴³ If metabolic activity increases cyclin D expression and, in turn, cyclin D inhibits metabolism, a rise in cyclin D in G1 would shut down metabolism, thereby preventing re-entry into G1 until the G1/S transition is complete.

pRB-E2F

Cyclin D-CDK4/6 complexes regulate the phosphorylation state of the retinoblastoma tumor suppressor (pRB) in G1; its phosphorylation abolishes inhibition of physically associated E2F/DP transcription factors, thereby allowing the expression of genes required for DNA synthesis.^44,45 Thus, phosphorylation of pRB in late G1 by cyclin D-CDK4/6 converts E2F/DP from transcriptional repressor into activator, thereby promoting S-phase entry (Fig. 2).

Given that E2F transcription factors are involved in cell cycle progression and survival, it is not surprising that roles for increased E2F activity have been shown in cancer.⁴⁵ It is well established that E2F promotes (cancer) cell proliferation through transcriptional activation of genes involved in cell cycle regulation and DNA synthesis. But a role for E2F in metabolism has also emerged. E2F1, as well as its upstream activator CDK4, has been shown to regulate adipogenesis by positively regulating PPARγ.^46,47 In accordance with this, pRB represses PPARγ at the early stages of adipocyte differentiation.⁴⁸ On the other hand, pRB acts positively on terminal adipocyte differentiation by binding directly to the transcription factor CCAAT/enhancer binding protein (C/EBP)β, thereby facilitating its transactivation.⁴⁹

Besides adipogenesis, the E2F-pRB pathway is involved in the glucose metabolism of pancreatic β-cells. E2F1-depleted mice display a diabetic phenotype, resulting from glucose intolerance and deficient insulin secretion.⁵⁰ In line with this, E2F1 regulates Kir6.2, a key component of the K_ATP channel controlling glucose-induced insulin secretion (Fig. 2).⁵¹ Expression of Kir6.2 is lost in the pancreas of E2F1-depleted mice, resulting in insulin secretion defects. Furthermore, E2F1 was shown to induce expression of pyruvate dehydrogenase kinase 4 (PDK4). PDK4 is one of 4 kinases that inhibit pyruvate dehydrogenase (PDH) and thereby oxidative metabolism (Fig. 2).⁵² Likewise, inactivation of pRB by the oncogenic E1A adenoviral protein triggers PDK4 expression (Fig. 2). Induction of PDK4 by pRB-E2F1 in myoblasts diverts pyruvate away from mitochondrial oxidation, thereby increasing lactate production and flux into biosynthetic pathways.⁵³ Furthermore, E2F1 stimulates glycolysis by upregulating the expression of the phosphofructokinase 2 PFK2 enzyme (Fig. 2).⁵⁴ Depletion of E2F1 increases the expression of regulators of mitochondrial biogenesis and function, such as mitochondrial topoisomerase I (TopImt).⁵⁵ Correspondingly, E2F1 downregulates oxidative and increases glycolytic genes expression.⁵⁶ In this way, the pRB-E2F1 pathway promotes a switch from oxidative to glycolytic metabolism, thereby supporting the metabolic phenotype typical of proliferating cells.

Altogether, the pRB-E2F pathway provides yet another mechanism for the coordinated regulation of metabolism throughout the cell cycle by facilitating glycolysis and repressing oxidative phosphorylation (Fig. 2). pRB-E2F-mediated stimulation of adipogenesis and inhibition of oxidative metabolism are in line with the observations made for cyclin D and suggest that the latter acts on these metabolic pathways at least to a significant extent through pRB-E2F. However, while pRB-E2F promotes glycolysis, cyclin D suppresses it. This implies that cyclin D acts on glycolytic enzymes in a pRB-E2F-independent manner.

CKIs

As cyclin-CDK complexes regulate metabolism, and because the activities and functions of CDK-cyclin complexes are regulated by the CKIs, the latter should also be associated with the cellular metabolism. Indeed, p21^Cip1 and p27^Kip1 were demonstrated to modulate adipocyte differentiation, since loss of either kinase inhibitor in mice induces adipocyte hyperplasia.⁵⁷ In accordance with this, combined disruption of p21^Cip1 and p27^Kip1 in mice induces an increase in the number of adipocytes and the development of hypercholesterolemia, glucose intolerance and insulin insensitivity, which are features of obesity. However, these observations require further investigation, as other studies have shown that p21^Cip1 null mice are less prone to obesity induced by lipid-rich diet, whereas p27^Kip1 null mice do not show hyperplasia of adipocytes but increased insulin secretion that prevents hyperglycemia in diabetic mice models.^58-60

The UPS and glycolysis

Cyclins and CKIs are both subject to tight temporal control and degradation by the UPS. Two ubiquitin ligases are crucial in the cell cycle. The ligase anaphase-promoting complex or cyclosome (APC/C) regulates both the transition through G1 and the exit from M phase by degrading both S- and M-phase cyclins (cyclin A and B) and Securin (allowing chromosome separation). Directing APC/C to correct substrates at specific time points in the cell cycle depends on one of 2 activators, CDC20 or CDC20 homolog (CDH1).^61,62 The latter is expressed and active in mitosis, until it is replaced by CDH1 in late mitosis and the G1 phase (Figs. 1 and 2).⁶³ The ligase Skp1/cullin/F-box protein (SCF) complex controls the G1/S-phase and the G2/M transition. The F-box protein of the SCF complex regulates its substrate recognition.^64,65 SCF-Skp2 mainly ubiquitinates and degrades CKIs such as p27^Kip1 and p21^Cip1, but also cyclin E, whereas SCF-β-TrCP positively regulates APC/C^CDC20 and CDK1 to ensure G2/M transition (Fig. 2). APC/C and SCF control each other to regulate progression of the cell cycle.⁶⁶

The yeast F-box protein GRR1, a component of SCF complex regulating G1/S transition, was first identified as being essential for adaptation to nutrient availability. In response to extracellular glucose, SCF^GRR1 blocks exit from the cell cycle and sporulation by targeting Ime2p kinase.⁶⁷ At the same time, SCF^GRR1 represses genes required for the utilization of alternative carbon sources, and upregulates hexose transporters.^68-71 On the other hand, when glucose is removed, SCF^GRR1-mediated degradation of Ime2p is abrogated, cells exit the cell cycle and sporulation proceeds.⁶⁷ Moreover, following glucose deprivation, SCF^GRR1 inhibits glycolysis by degrading Pfk27, the yeast homolog of the glycolytic enzyme PFKFB3 (Fig. 2).²⁸ Thus, SCF^GRR1 in yeasts regulates both cell cycle exit and metabolism as a function of glucose availability.

Shortly after Pfk27 was identified as the target of SCF^GRR1 in yeast, PFKFB3 was found to be a degraded by APC/C^CDH1 in neurons (Fig. 2).²⁹ Constitutive breakdown of PFKFB3 helps terminally differentiated cortical neurons to preserve their low glycolytic state. It also prevents oxidative damage by redirecting glucose-derived metabolites into antioxidants-providing pentose phosphate pathway (PPP). In dividing cells, PFKFB3 is directed for degradation both by APC/C^CDH1 during late mitosis and G1 and subsequently by SCF^β-TrCP in late S phase (Figs. 1 and 2).^72-75 Overexpression of the APC/C activator CDH1 leads to degradation of PFKFB3 and thereby restricts glycolysis. Accordingly, depletion of CDH1 promotes glycolysis in a PFKFB3-dependent manner and stimulates cells to enter S phase.^73,74 Inactivation of APC/C^CDH1, resulting from a phosphorylation of CDH1 that occurs in late G1, leads to the accumulation of PFKFB3 and consequently promotes both proliferation and glycolysis (Fig. 1). In late S-phase, PFKFB3 levels drop again due to an increase in activity of SCF^β-TrCP, which specifically directs PFKFB3 for degradation (Fig. 1).^72,75 Restricted expression of PFKFB3 to the specific window of late G1 and early S phase generates a peak in anaerobic glycolysis during the G1/S transition and a PPP peak in S phase (Fig. 1, upper panel). Thus, the joint action of the APC/C and SCF complexes on PFKFB3 coordinates metabolic activity and cell cycle progression.

APC/C^CDH1 was recently shown to also regulate phosphatase and transactivator EYA1.⁷⁶ The level of EYA1 protein oscillates in the cell cycle, peaking during mitosis and dropping radically as cells enter G1, when APC/C^CDH1 reaches its peak level (Figs. 1 and 2). While depletion of CDH1 stabilizes the EYA1 protein, overexpression of CDH1 reduces its levels. Thus, APC/C^CDH1 degrades EYA1 precisely during M/G1 transition. EYA1 is required for proliferation during embryogenesis^77-79 and its level is elevated in several cancers.^80-83 Interestingly, EYA1 is also known to reprogram the aerobic metabolism of slow-twitch muscle fibers, which depend on lipid oxidation, into the glycolytic phenotype of fast-twitch muscle fibers, which depend mostly on glycogen as an energy source. This is achieved through the upregulation of the glycolytic enzymes aldolase A and β-enolase (Fig. 2).⁸⁴ Although only correlative, this might provide another mechanism by which UPS regulates glycolysis.

The UPS and glutaminolysis

Besides PFKFB3, the APC/C^CDH1 also directs the glutaminolytic enzyme glutaminase 1 (GLS1) for degradation during mitotic exit and G1 (Figs. 1 and 2).^74,75 Glutamine is important for proliferation as it replenishes tricarboxylic acid (TCA) cycle intermediates used for macrosyntheses of amino acids, lipids and nucleotides. GLS1 levels and the glutaminolysis rate rise after APC/C^CDH1 activity declines in late G1. Accordingly, depletion of APC/C activator CDH1 leads to an increase in cellular GLS1 concentration and consequently glutamine metabolism. In contrast to PFKFB3, GLS1 is not a substrate for SCF and therefore it is not degraded through the S and G2 phases, but only when the cells progress to the G2/M transition (Figs. 1 and 2). The distinct regulation of PFKFB3 and GLS1 proteins and therefore glycolysis and glutaminolysis, suggests the different functions of glucose and glutamine at particular phases of the cell cycle. Indeed, studies in synchronized cells show that both glucose and glutamine are required throughout G1, whereas only glutamine is needed to progress through S phase into the cell division phase.⁷⁵ The importance of regulating GLS1 for cell proliferation is further supported by the notion that GLS1 C, an isoform that is not targeted by the APC/C^CDH1, is overexpressed in several tumors.⁸⁵

Taken together, while ubiquitination by the UPS was originally recognized for its major role in regulating the cell cycle machinery, it is now also acknowledged for its integrated regulation of metabolism and proliferation. The dual regulation by these 2 ubiquitin ligases, APC/C and the SCF complex, delineates a differential regulation of both aerobic glycolysis and glutaminolysis during distinct phases of the cell cycle. In this sense, the UPS plays a major role in the provision of a specific metabolic profile reminiscent of proliferating (cancer) cells, including the Warburg effect, upregulation of the PPP and increased glutamine utilization.

p53 family members

p53 is a major tumor suppressor protein, which is mutated in many types of cancer. The activation of p53 in response to a plethora of stress signals provides the cell with 2 options: inhibition of the cell cycle at G1/S by inducing transcription of the CKI p21^Cip1, or induction of pro-apoptotic signals.⁸⁶ This allows cells to either repair the damage before engaging in cellular division, or, if the damage is beyond repair, to prevent cells from proliferating ever again. But recent research shows that p53 acts also as a major regulator of cellular metabolism.

Several functions of p53 have been demonstrated to silence glycolysis and promote oxidative metabolism. It hampers glycolytic flux through the various steps of glycolytic pathway by downregulating the expression of glucose transporters GLUT 1 and GLUT4, decreasing levels of the glycolytic enzyme phosphoglycerate mutase (PGM).^87,88 It also induces TIGAR, an enzyme that lowers the level of glycolytic activator fructose-2,6-bisphosphate (F-2,6-BP) and thereby inhibits glycolysis (Fig. 2).⁸⁹ Moreover, p53-mediated repression of monocarboxylate transporter 1 (MCT1) expression prevents the secretion of lactate under anaerobic conditions, which also reduces glycolysis.⁹⁰ Inhibition of the glycolytic pathway would be expected to be associated with redirecting glucose-derived metabolites into PPP pathway. However, p53 evades this by binding and inhibiting the rate-limiting enzyme of PPP, glucose-6-phosphate dehydrogenase (G6PDH; Fig. 2).⁹¹ p53 also downregulates the expression of the ribonucleotide reductase subunit M2 (RRM2), an enzyme controlling dNTP synthesis (Fig. 2).⁹² Inhibition of glycolysis and PPP is paralleled by a role of p53 in promoting oxidative metabolism. Next to maintaining mitochondrial balance,⁹³ p53 stimulates oxidative phosphorylation by upregulating cytochrome c oxidase (mammalian COX/yeast SCO2), a component of the mitochondrial electron transport chain (Fig. 2).^94,95 Moreover, p53 increases the TCA cycle rate by downregulating PDH-inhibitory kinase PDK2 and TCA cycle - associated malic enzymes (MEs; Fig. 2).^96,97 Additionally, p53 increases the expression of glutaminolytic enzyme glutaminase 2 (GLS2; Fig. 2).⁹⁸ GLS2 converts glutamine to glutamate, which can feed the TCA cycle, but can also participate in glutathione synthesis to control the redox state of the cell. Thus, by upregulating GLS2, p53 influences both glutamine metabolism and redox status. p53 also regulates the cellular redox state upon serine withdrawal. It does so by allowing de novo serine to be channeled to glutathione synthesis, which occurs at the cost of nucleotide synthesis.⁹⁹ Finally, p53 was shown to negatively regulate lipogenesis (Fig. 2). p53 inhibits fatty acid (FA) synthesis in mouse adipose tissue by suppressing the expression levels of transcription factor sterol regulatory element-binding protein (SREBP1c) and FASN and ATP citrate lyase (ACLY) (Fig. 2).¹⁰⁰ At the same time, p53 stimulates FA oxidation by inducing the expression of carnitine palmitoyltransferase (CPT), which is responsible for the transport of FA into the mitochondria, and guanidinoacetate methyltransferase (GAMT), an enzyme involved in creatine synthesis (Fig. 2).^101,102 Hence, p53 promotes cell cycle arrest not only by acting on the cell cycle machinery, but also by counteracting the general metabolic profiles favorable for proliferation and by supporting oxidative metabolic reactions characteristic of resting cells.

The other p53 family members p63 and p73, too, have metabolic functions, broadening the impact of the p53 family on cell metabolism. The Tp63 and Tp73 genes are transcribed from 2 separate promoters, encoding either full- length proteins that retain a full transactivation (TA) domain (TAp63 and TAp73) or N-terminally truncated isoforms (ΔNp63 and ΔNp73). Both TAp63 and TAp73 upregulate the expression of GLS2 and thereby increase glutaminolysis.^103,104 Moreover, TAp73 stimulates oxidative metabolism by upregulating COX subunit 4¹⁰⁵. Correspondingly, the depletion of TAp73 results in a decrease in both respiration and ATP production. In contrast to p53, TAp73 activates the expression G6PDH and the flux to PPP, therefore redirecting glucose for the synthesis of nucleotides and antioxidants.¹⁰⁶ TAp73 also regulates amino acid metabolism by increasing the levels of serine, glycine, and glutathione.¹⁰⁷ Similarly to p53, p63 was implemented in the regulation of lipid metabolism. Loss of TAp63 disrupts lipogenesis, FA synthesis and oxidation, and protects against insulin resistance in mice.¹⁰⁸ Likewise, ΔNp63 transcriptionally activates FASN.¹⁰⁹ Altogether, the p53 family provides yet another element contributing to direct crosstalk between the cell cycle machinery and cellular metabolism.

Metabolic Regulation of the Cell Cycle Machinery

While it is important for a cell to provide sufficient building blocks to enable cell division, this mechanism also functions vice versa: it is equally important for a cell to adapt its cell cycle to the environment (i.e., changes in nutrient availability) and the state of metabolism. As early as in 1974 it was shown that cells in the absence of glucose arrest at the G1/S restriction point¹¹⁰. This demonstrated that glucose availability governs as a metabolic cell cycle checkpoint. Since then, various mechanism through which cells synchronize their cell cycle with their metabolic state have been discovered, several of which are described below.

PFKFB3

Beyond its metabolic activity, PFKFB3 has been described to localize in the nucleus and regulate the cell cycle machinery.¹¹¹ Overexpression of nuclear PFKFB3 is accompanied by increased expression of G1-promoting cyclin D3 (Fig. 2). Moreover, nuclear PFKFB3 stimulates proliferation by increasing the expression of mitotic kinase CDK1 and M phase-promoting phosphatase Cdc25C and by decreasing the expression of the CDK1 inhibitor p27^Kip1 (Fig. 2). These effects are completely abrogated by mutating either the active site or nuclear localization residues of PFKFB3, demonstrating a requirement for nuclear delivery of F-2,6-BP in this setting. Consequently, addition of F-2,6-BP to cell lysates promotes CDK1‐mediated p27^Kip1 phosphorylation, which is a cue for p27^Kip1 degradation. Along these lines, suppression of PFKFB3 induces cell cycle delay.⁷² These data show that, while PFKFB3 is subject to a tight regulation by the UPS, it can return this action by regulating the cell cycle machinery.

PKM2

PKs catalyze the final rate-limiting step of glycolysis, generating ATP and pyruvate. PK isoform M2 (PKM2) is specifically enriched in highly proliferating cells, including cancer cells, where it regulates aerobic glycolysis.¹¹² Apart from its important metabolic function, PKM2 also has a non-metabolic role in the control of the cell cycle progression. In response to epidermal growth factor (EGF) stimulation, PKM2 translocates to the nucleus where it binds to β-catenin and promotes its transcriptional activity.^113,114 PKM2-β-catenin complex subsequently localizes to the cyclin D1 (CCND1) and c-Myc promoters and enhances their expression through detachment of a histone deacetylase, HDAC3.¹¹⁵ Thus, by inducing cyclin D1 expression, PKM2 regulates the G1-S phase transition (Fig. 2). Besides PKM2s role in regulating the expression of cyclin D1, its direct involvement in the regulation of cell cycle progression has also been demonstrated. During mitosis, PKM2 binds to and phosphorylates the spindle assembly checkpoint (SAC) protein Bub3 (Fig. 2).¹¹⁶ This phosphorylation is required for the formation of Bub3-Bub1-Blinkin complex and correct kinetochore microtubules attachment, allowing for proper chromosomal segregation. In this way, PKM2 regulates not only G1/S transition but also progression through mitosis.

Acetyl-CoA

Cytosolic and nuclear acetyl-CoA is not only an important intermediate for macrosyntheses, but also a precursor for the posttranslational modification of proteins by acetylation. For example, acetylation of histones is dependent on acetyl-CoA-producing ACLY. In various mammalian cell types, depletion of ACLY, and therefore a drop in acetyl-CoA levels, decreases histone acetylation.¹¹⁷ On the other hand, in yeast, a rise in acetyl-CoA upon depletion of ACC increases the acetylation of histones.¹¹⁸ Notably, acetylation of histones is an essential process for the release of DNA for replication and therefore for cell cycle progression.^119,120 Thus, by regulating the acetylation of histones, acetyl-CoA controls the cell cycle (Fig. 2).

In addition to acetylation, availability of acetyl-CoA regulates the glycosylation of proteins (Fig. 2). O-linked N-acetylglucosamine (O-GlcNAc) modifications depend on UDP-GlcNAc, production of which is controlled by the availability of glucose, glutamine and acetyl-CoA.¹²¹ O-GlcNAc-transferases (OGTs) transfer UDP-GlcNAc onto a target protein, while O-GlcNAcases (OGAs) remove it, permitting a dynamic regulation of O-GlcNAcylation levels. In recent years, numerous proteins have been identified as substrates for O­GlcNAc modification, including regulators of cell cycle progression like p53.¹²² Moreover, O-GlcNAcylation levels have been found to vary along the cell cycle (Fig. 1). Serum addition triggers G0/G1 transition by activating PI3K/AKT and MAPK pathways and transcription of cyclin D1. OGT levels are significantly increased following serum stimulation,¹²³ while OGT depletion delays G1 entry and prevents serum-induced cyclin D1 synthesis (Fig. 2).¹²⁴ Hence, OGT and therefore increased O-GlcNAcylation, are vital for entry into the cell cycle. Contrary to that, at the G1/S transition, global O-GlcNAcylation is decreased due to elevated OGA activity.¹²³ Interestingly, this leads to lower O-GlcNAcylation of histones, which, similarly to the histone acetylation mentioned above, promotes the relaxation of DNA, necessary for replication and cell cycle progression.^125,126 At the G2/M checkpoint, O-GlcNAcylation reaches its peak.^127-130 Accordingly, depletion of OGT decreases cyclin B1 expression and therefore impairs G2/M transition (Fig. 2).¹²⁷ Although we are only beginning to understand the importance of O-GlcNAcylation in regulating cell proliferation, the evidence discussed provides additional indications that the cell cycle machinery and metabolism are tightly intertwined.

ATP/AMP ratio

The ATP/AMP ratio reflects the energy status of the cell. The protein complex that plays a crucial role in regulating cellular energy is AMP-activated protein kinase (AMPK).¹³¹ Under conditions of energetic stress such as glucose deprivation and low ATP levels, activated AMPK negatively regulates energy-consuming metabolic processes such as protein and lipid synthesis, while at the same time it promotes energy-producing oxidative phosphorylation and FA oxidation (Fig. 2).¹³² It has been recently suggested that AMPK can regulate energy levels also by direct control of the cell cycle machinery. Along these lines, activation of AMPK upon glucose deprivation or treatment with AMP analog AICAR causes an arrest in the G1 phase.^133,134 This cessation of proliferation is associated with phosphorylation of p53 at Ser15 and upregulation of CKI p21^Cip1 expression, a target of activated p53 (Fig. 2). AMPK phosphorylates also the C-terminal residue of another CKI, p27^Kip1, causing its stabilization (Fig. 2).¹³⁵ Likewise, methylene blue (MB)-mediated activation of AMPK reduces expression of cyclins A2, B1 and D1, leading to proliferative arrest.¹³⁶ Moreover, activation of AMPK due to mitochondrial dysfunction promotes p53-dependent transcription of the F‐box protein archipelago. Archipelago then recruits cyclin E to the SCF complex, resulting in cyclin E degradation and G1-S cell cycle arrest (Fig. 2).^137,138 As such, AMPK plays a pivotal role as a metabolic cell cycle checkpoint, preventing cell cycle entry in conditions of low nutrient availability.

NAD⁺/NADH ratio

The NAD⁺/ NADH ratio has an important role in the regulation of redox homeostasis stress and is often considered to be a readout of the metabolic status. NAD⁺ is converted to NADH in catabolic reactions including glycolysis and the TCA cycle. To maintain a proper redox state, NADH is regenerated constantly via several mechanisms, such as oxidation in the mitochondrial respiratory chain and reduction of pyruvate to lactate in the last step of glycolysis.¹³⁹ NAD⁺ is a classical coenzyme mediating redox reactions,¹³⁹ but also plays an important role in regulation of NAD⁺ - consuming enzymes, including sirtuin family of NAD⁺-dependent deacetylases.¹⁴⁰ Notably, in recent years, several mechanisms linking sirtuins to the cell cycle machinery have been described. SIRT2 controls mitotic exit and acts as a checkpoint protein in cells treated with microtubule poisons.^141,142 Under genotoxic stress, SIRT1 deacetylates and hinders the activity of p53, thereby preventing p53-mediated transactivation of p21^Cip1 and cell cycle arrest (Fig. 2).^143,144 SIRT1 also regulates the cell cycle by deacetylation of the forkhead box O 3 (FOXO3) transcription factor.¹⁴⁵ In response to oxidative stress, SIRT1 binds to and deacetylates FOXO3, resulting in increased levels of FOXO3 target p27^Kip1 and induction of cell cycle arrest (Fig. 2). Moreover, SIRT1 has been described to regulate the components of the circadian clock machinery.^146,147 Interestingly, among clock-controlled genes are those that have an essential role in cell cycle control, including cyclin D1 and inhibitor of CDK1-cyclin B1 complex, Wee1.^148,149 Hence, by regulating the circadian clock, SIRT1 inhibits cells cycle progression in situations of stress (Fig. 2). Taken together, sirtuins translate the NAD+/NADH ratio to the components of the cell cycle machinery.

Concluding Remarks

The combined regulation of metabolic events and components of the cell cycle machinery described here delineates an intricate relationship between the 2. Their reciprocal regulation plays a pivotal role in the cellular decision to enter and progress through cell cycle and to adjust metabolic pathways in response to both intracellular and external cues. Cells strictly control the cell cycle machinery components, in order to produce the required macromolecules during specific stages of the cell cycle. However, sometimes the regulation is disturbed, which leads to severe pathologies, including cancer. Thus, studying the reciprocal regulatory networks linking the cell cycle and metabolism is likely to reveal potential targets for therapy.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This work was supported by a Vici grant from the Netherlands Organization for Scientific Research (NWO) and a Queen Wilhelmina Award grant from the Dutch Cancer Society (KWF Kankerbestrijding) to DSP.