Role of purines in regulation of metabolic reprogramming

Abstract

Purines, among most influential molecules, are reported to have essential biological function by regulating various cell types. A large number of studies have led to the discovery of many biological functions of the purine nucleotides such as ATP, ADP, and adenosine, as signaling molecules that engage G protein-coupled or ligand-gated ion channel receptors. The role of purines in the regulation of cellular functions at the gene or protein level has been well documented. With the advances in multiomics, including those from metabolomic and bioinformatic analyses, metabolic reprogramming was identified as a key mechanism involved in the regulation of cellular function under physiological or pathological conditions. Recent studies suggest that purines or purine-derived products contribute to important regulatory functions in many fundamental biological and pathological processes related to metabolic reprogramming. Therefore, this review summarizes the role and potential mechanism of purines in the regulation of metabolic reprogramming. In particular, the molecular mechanisms of extracellular purine- and intracellular purine-mediated metabolic regulation in various cells during disease development are discussed. In summary, our review provides an extensive resource for studying the regulatory role of purines in metabolic reprogramming and sheds light on the utilization of the corresponding peptides or proteins for disease diagnosis and therapy.

Background

In 1924, a sharply distinct metabolism of tumor cells, the enhancement of glycolysis and decrease of the oxidative phosphorylation (OXPHOS), was firstly reported by Otto Warburg [1–3]. This type of tumor cell-specific metabolism favoring anaerobic over aerobic metabolism occurred even if environmental oxygen supply is abundant [1, 2, 4, 5]. This phenomenon was observed in different types of tumor cells and was later named the “Warburg effect” [6–9]. The mainstream view that has been established suggests that metabolic alterations contribute to the proliferation of tumor cells and are strictly programmed by a series of complicated pathways that remain unknown [4, 7, 10, 11]. Therefore, similar to transcriptional reprogramming, the concept of “metabolic reprogramming” was formed. Furthermore, the molecular characteristics of metabolic reprogramming have also been expanded to include changes in the activity of certain key enzymes in metabolic pathways, such as lipolysis and PPP, and metabolic regulators, including transporters of glucose or amino acids [12–14].

Recent studies confirmed that metabolic reprogramming is a key mechanism involved in physiological and pathological processes in cells via regulation of proliferation, differentiation, activation, and death [15–20]. Although it has been shown that metabolic reprogramming has an essential impact on cellular function, the molecular mechanism underlying the regulation of metabolic reprogramming remains unclear. This review summarizes the regulatory role of purines involved in metabolic reprogramming and in molecular mechanisms and provides a systematic perspective for further translational practice (Table 1).

Table. 1.

The current realm of metabolic reprogramming

Cell type	Biological process	Metabolic alteration	Reference
Solid tumor cell	Proliferation, survival, and oncogenesis	Glycolysis, OXPHOS, PPP, glutaminolysis, and lipolysis	[7, 11]
Leukemia	Cell growth and autophagy	Glutaminolysis, glycolysis, and lipid metabolism	[8, 9]
Macrophages	Macrophages polarization	Glycolysis, l-arginine metabolism, glutaminolysis, OXPHOS, and lipolysis	[21, 22]
T cells	Activation and differentiation	Glycolysis, glutaminolysis, FAO, and PPP	[21, 23]
B cells	Activation, differentiation, and antibody production	Glycolysis, OXPHOS, and PPP	[21, 22, 24]
Dendritic cells	Activation	OXPHOS and glycolysis	[21]
MDSCs	Activation, differentiation, and inflammatory response	Amino acid metabolism, glycolysis, and FAO	[21, 25]
Neutrophils	Microbicidal functions	Glycolysis and PPP	[21]
Natural killer cells	Activation and killing function	Glycolysis and OXPHOS	[26]
Induced pluripotent stem cells	Generation of iPSCs from somatic cells	Glycolysis and OXPHOS	[27]
Mesenchymal stem cells (MSC)	Survival and activation	glycogen synthesis and glycolysis	[19, 28]
Hematopoietic stem cells (HSCs)	Lineage commitment, proliferation	Glucose metabolism, glutamine metabolism, FAO	[19, 29]
Erythrocytes	Oxygen release	Glycolysis	[30, 31]

OXPHOS, oxidative phosphorylation; PPP, pentose phosphate pathway; FAO, fatty acids oxidation; MDSCs, myeloid-derived suppressor cells

The growing realm of metabolic reprogramming

The realm of metabolic reprogramming has been expanded to include not only tumor cells but also different, active cell types, including immune cells and stem cells, due to the development of omics such as proteomics, metabolomics, and metabolite tracking, including ¹⁸F-deoxyglucose positron emission tomography (FDG-PET) and metabolic flux analysis [6, 12, 19, 21, 29]. Considerable evidence has proven that metabolic reprogramming can regulate cell fate or commitment of stem cells via epigenetic regulation by metabolites, possibly contributing to the regenerative ability of stem cells [19, 27–29, 32–34]. The metabolic state of the organism affects the immune system through an altered fuel supply and thus the cellular metabolism used by immune cells. These affected components, in turn, regulate immune cell function, which we now know can subsequently regulate whole body metabolism such that an immunometabolic loop exists between systemic and cellular metabolism in the organism [35]. It has been reported that enhanced glycolysis and PPP are essential for the activation of T cells [21, 36]. Other immune cells, including B cells, macrophages, dendrite cells, and natural killer (NK) cells, play roles in the immune response with observed metabolic reprogramming [21–24, 26]. Of note, through regulation of the activation of immune regulatory cells such as regulatory T (Treg) cells and myeloid-derived suppressor cells (MDSCs), metabolic reprogramming might play a critical role in the self-regulation of the immune system [21, 25, 37, 38]. In addition, modulation of metabolic reprogramming for the treatments of autoimmune diseases and chronic inflammation has also increased expectations, especially for treatments of T cell–related disease [39, 40]. For tumor immunotherapy, metabolic rewiring has been reported to regulate therapeutic outcomes by significantly modulating tumor cells, tumor-associated immune cells, and tumor microenvironments [21, 41]. Several drugs targeting metabolic alterations have been utilized in preclinical or clinical trials [42]. For instance, metabolic reprogramming and current targets, such as programmed cell death protein-1 (PD-1), are being reevaluated. Studies have confirmed that metabolic alteration in fatty acid oxidation (FAO) or glycolysis is triggered by PD-1 ligation [43]. Additionally, hypoxia-induced factor (HIF), which is critically involved in metabolism, contributes to the induction of programmed cell death-ligand 1 (PD-L1) [44]. Therefore, metabolic reprogramming might start a revolution in the classic system of tumor pharmacology. Of note, our group discovered the critical role of metabolic reprogramming in erythrocyte function and hypoxia response, expanding the vision for metabolic reprogramming to include nonproliferating cells [30, 45]. Therefore, manipulating metabolic reprogramming could further help us develop novel targets for disease treatments.

The mechanism of metabolic reprogramming–mediated regulation of cellular processes largely depends on the sequential metabolites that alter metabolism. Current studies have revealed that small metabolites can be released into the extracellular microenvironment and function as ligands, for example, purines that bind to a matched receptor [30]. In addition, metabolic reprogramming provides energy and material pools that enable functional change through increased production of ATP and nucleic acid.

Currently, thousands of articles aimed at metabolic reprogramming are being published, greatly enriching the theoretical foundation of metabolic reprogramming. Undoubtedly, the concept and realm of “metabolic reprogramming” are still developing.

Overview of purines

The metabolism of purines

Biological purines are a series of small molecules with purine rings that exist inside or outside cells. Among all kinds of purines, the most frequently mentioned is probably adenosine-5′-triphosphate (ATP), which has been recognized as a critical metabolic molecule within the cell and was first described as an extracellular messenger by Burnstock in 1972 [46, 47]. Under normal conditions, ATP is produced from the gradual phosphorylation of ADP/adenosine monophosphate (AMP), forming a primary energy machine for biological processes [47]. Serving as the raw purinergic material of ADP/AMP/ATP, adenosine can be bypass degraded with the help of several enzymes including adenosine deaminase (ADA), S-adenosylhomocysteine hydrolase (SAHH), and adenosine kinase (ADK), leading to the generation of inosine, hypoxanthine, adenosylhomocysteine (AdoHcy), and AMP [48]. Additionally, intracellular purines can be utilized for the generation of cyclic purines such as cAMP, which is among the most famous second messengers [49].

Under stress, however, ATP tends to be released from injured or necrotic cells and subsequently degraded by four ectonucleotidase groups, namely, ectonucleoside triphosphate diphosphohydrolases (CD39), ecto-59-nucleotidase (CD73), ectonucleotide pyrophosphatase/phosphodiesterase (NPPs), and alkaline phosphatases, eventually leading to adenosine production, which serves as the major source of extracellular purines [50]. The other sources of extracellular adenosine involve direct secretion with the help of equilibrative nucleoside transporters (ENTs), which enable adenosine to travel bidirectionally [51]. Some of these extracellular purines, including ATP and adenosine, function as signal molecules and bind to exclusive receptors to complete their various regulatory functions [48].

The biological function of purines

The components and regulatory process of purines in cytoplasm differ from those in the extracellular microenvironment; therefore, herein, we elaborate on their different biological functions.

Extracellular purines

Extracellular purines mainly regulate cellular functions as a signaling ligand by engaging receptors on the cell surface. Purine receptors can be divided into two groups: P1 receptors for adenosine binding and P2 receptors for other pyrimidine nucleotide binding. The distribution of these receptors varies depending on different cell types, which determines the complex downstream regulatory network of the purinergic pathway [52]. Additionally, some of these receptors are coupled with G proteins and are capable of generating second messengers such as cyclic purines [53].

Adenosine signaling involves adenosine and four receptors and has a large impact on physiological homeostasis in many ways [48]. Under pathological conditions, the concentration of extracellular adenosine can elevate dramatically in response to stress or assist in repairing cell damage [48]. The adenosine receptor can be divided into four types (ADORA1, ADORA2A, ADORA2B, and ADORA3), which are all coupled with G proteins to function [54]. The role of the adenosine pathway may be tissue-specific because of the various distribution patterns of the adenosine receptor [48]. The adenosine pathway plays a beneficial role in cardiovascular protection; therefore, targeted drugs have been seriously investigated or already applied in clinical practice [55]. Adenosine-related research, in recent years, has focused on the cellular response to stress, hypoxia, immunity, inflammation, etc. [48, 55, 56].

ATP receptors can be subclassified according to their transduction mechanism (i.e., as metabotropic P2YRs or ionotropic P2XRs) [57]. The fast-acting P2XRs constitute a family of seven ligand-gated ion channel receptors (P2X1–7) that have a single physiological agonist (ATP). Composed of subunits of 379–595 amino acids in length, they share a similar tertiary topology with an intracellular NH2 and a more extended COOH terminus, a large extracellular loop responsible for ligand binding, and two transmembrane-spanning regions (TM1 and TM2) [58, 59]. Two hydrophobic transmembrane domains (M1 and M2) are separated by an ectodomain within which 10 cysteine residues form disulfide bridges [59]. TM1 is involved in channel gating, and a helix of TM2 forms a channel pore. Following ATP binding, P2XRs become permeable to Na⁺, K⁺, and Ca²⁺. Moreover, activation of a P2X7R induces formation of a large pore that enables the passage of molecules as large as 900 Da [52]. Unlike the P2XRs, P2Yrs belong to the G protein–coupled receptors and are not accessible exclusively to ATP. In fact, P2Yrs can be selectively engaged and activated by small metabolites including ATP, ADP, and UDP [60]. Based on current understanding, the main role of P2Yrs is thought to be involved in cardiovascular regulation and immunological regulation, where they are considered to be a potential drug target [61]. However, little research has been carried out.

In general, there are other types of purines that exist extracellularly, such as inosine and hypoxanthine, which mainly originate from the degradation of extracellular adenosine. Studies have shown nonexclusive binding between inosine and several types of receptors, such as GABA_A receptors and adenosine receptors [62, 63]; however, little is known about the exclusive receptors for or the signaling pathways of inosine or hypoxanthine. Since little evidence has been generated to show an obvious relation between metabolic reprogramming and inosine or hypoxanthine, the extracellular purines discussed in this review mainly refer to ATP and adenosine.

Intracellular purines

In addition to extracellular purines, intracellular purines, which are mainly classified as either noncyclic purines or cyclic purines, have a vital and common regulatory role. There are some molecular sensors for intracellular purines, and their purine-dependent activation can initiate a signal transduction cascade.

Noncyclic purines exist both inside and outside cells and include adenosine, AMP, ADP, and ATP. Apart from the extracellular effect on receptors, AMP and ADP, in the past decade, were reported to regulate metabolism via their intracellular sensor, AMP-activated protein kinase (AMPK). During energy-depleting exercise, an increase of free AMP and ADP, which can directly bind to the γ regulatory subunits of AMPK, leads to a conformational change that activates AMPK via phosphorylation. The process can be elicited by even modest decreases in ATP production that result in relative increases in AMP or ADP [64]. In a coordinated response, AMPK increases metabolic output through nutrient utilization in addition to inhibiting anabolic pathways, thus enabling adaptive changes in growth, differentiation, and metabolism during energy deficiency [65]. In mammals, AMPK plays a general role in coordinating growth and metabolism and special roles in regulating metabolism in dedicated tissues such as fat, neuron, and muscle [66].

Small intracellular cyclic purines, including cyclic AMP (cAMP) and cyclic GMP (cGMP), have been well recognized as critical second messengers within cells. As the first-observed second messenger, cAMP was proven to be engaged in various physiological functions including memory formation, muscle contraction, and endocrine regulation, such as insulin signaling [67–69]. The effect of cAMP largely depends on the activation of cAMP receptor proteins (CRPs) or protein kinase A (PKA), a sensor of the concentration of intracellular cAMP levels that functions in downstream signaling to regulate the biological reactions or genetic expression within the cell [49, 70]. Similar to cAMP, cGMP interacts with some intracellular molecules such as protein kinase G (PKG) to fulfill downstream action [71]. In addition, cyclic purines have been revealed as activators of transcription factors or cation channels, indicating an epigenetic regulatory role of intracellular purines [69].

The integration of the metabolized purines and the regulatory pathway mediated by purines is illustrated in Fig. 1.

Fig. 1.

An integration of pathways and metabolism of biological purines: intracellular ATP is partly metabolized by adenylate cyclase (AC) to cAMP which could activate PKA. Cells release ATP through pannexin channels under stress. The accumulation of extracellular ATP is dephosphorylated to adenosine by 4 ecto-nucleotidases including CD39, CD73, NPPs, and alkaline phosphatases. Adenosine can further be metabolized by adenosine deaminase (ADA) to inosine (I) and hypoxanthine (HypoX) or function as a signaling molecule by activation of its adenosine receptors on multiple cell types. Once uptake by equilibrative nucleoside transporters (ENTs), adenosine is further metabolized by ADA to inosine, S-adenosylhomocytesine hydrolase (SAHH) to adenosylhomocysteine (AdoHcy) or adenosine kinase (ADK) to AMP, which could stimulate AMPK when [AMP]/[ATP] ratio escalates

Purine-mediated regulation of metabolic reprogramming

According to the classification of purines described above, in this part, the regulation of metabolic reprogramming by extracellular and intracellular purines is summarized.

Extracellular purines

Adenosine pathway

As mentioned above, the adenosine pathway is involved in cellular functions in response to stress, hypoxia, immunity, and inflammation [48, 56, 59]. These response functions, however, link closely to metabolic changes within the cell, and metabolic reprogramming is common under these conditions. The regulatory role of adenosine in metabolic reprogramming within inflammatory or immune cells has been confirmed, including the positive effect of adenosine/ADORA2 signaling in macrophage activation via increased fructose 2,6-bisphosphate and glycolytic flux [22, 72, 73]. Moreover, a series of studies by Sitkovsky and coworkers have demonstrated that adenosine/ADORA2A signaling contributes greatly to the suppression of T cells according to the tumor microenvironment, which is significantly regulated by metabolic reprogramming [31, 74–76]. Since the adenosine pathway can upregulate lipolysis and mobilize energy stores to cope with caloric deficiency, there is a high probability that metabolic reprogramming regulated by adenosine signaling plays a leading role in the progression of tumor cells in energy-deprived environments [77–79]. On the other hand, researchers have attempted to explore how metabolic reprogramming regulates the development of B cells and dendritic cells via the adenosine pathway. They hypothesized that the kind of impact that the adenosine pathway has on metabolic reprogramming of immune cells may explain the complicated mechanism of immune therapy resistance [21, 80]. In addition to contributing to the well-known understanding of metabolic reprogramming in mitotic cells, our group found a novel role for hypoxia-triggered adenosine signaling in erythroid physiology [30, 81, 82]. Using a translation study design and mature metabolomics protocols, we successfully confirmed that adenosine/ADORA2B/AMPK signaling triggers metabolic reprogramming, further inducing 2,3-BPG production in erythrocytes and providing materials for nucleotide synthesis in erythroid progenitor cells, thereby enhancing oxygen release and stress hematopoiesis, respectively, during oxygen crisis [30, 82]. This discovery led to the prospect of purine as a drug target or biomarker for nonmitotic cellular diseases. Intriguingly, these facts and hypotheses are almost accompanied by microenvironmental hypoxia, which has been proven to activate the adenosine pathway and promote metabolic reprogramming. Indeed, serving as the key factor in hypoxia metabolism, HIF might act as the bridge between adenosine signaling and metabolic reprogramming. According to the classification and working mechanism of HIF, the variable stability of HIF-1α and HIF-2α determines the effect of oxygen availability on transcription, leading to a change in biological processes [83]. Initially, studies were focused on the upregulation of purinergic receptors mediated by HIF [56]. However, recent studies have shown that both HIF-1α and HIF-2α can function downstream of purinergic signaling under hypoxic conditions [84–86]. More attention has been paid to HIF-1α owing to its ubiquitous distribution, although HIF-2α is expressed distinctively among tissues [87]. In addition, HIF-1α is reportedly regulated reciprocally by the adenosine pathway, and the downstream factors of HIF include a series of molecules that regulate metabolism, such as glycolytic enzymes and those involved in glucose transport [56, 88–91]. Therefore, we suspect that HIF, along with the adenosine pathway, might function as an indispensable component for the regulation of metabolic reprogramming.

However, the relationships between the adenosine pathway and metabolic reprogramming cannot be simply unidirectional. Apart from the regulatory effect that adenosine has on metabolism, metabolic reprogramming can also contribute to the generation of adenosine and eventually promote the process of adenosine/receptor/metabolic reprogramming [92, 93]. This kind of feedback can greatly amplify the downstream effect. For example, the activated adenosine pathway has been verified as a promoter of the regeneration of pancreatic β cells by inducing glycolysis [94], which potentially adds to the generation of ATP and subsequent adenosine. Thus, targeting the adenosine pathway through metabolic reprogramming would be highly efficient for treatment approaches.

ATP pathway

Considering that the concentration of ATP in the extracellular milieu is extremely low under physiological conditions (in the nanomolar range), even a small amount of ATP release can elicit a relatively strong signal against the background, characterizing ATP-based purinergic signaling as an extracellular messenger system with a high signal-to-noise ratio [95].

The cradle of knowledge about purinergic signaling was neurotransmission, for which ATP was long known to have a well-defined role as a neuromediator. During recent decades, our knowledge of how ATP signaling impacts brain pathology has increased considerably [96]. During a mild nonlethal ischemic episode, astrocytes exert neuroprotective actions via the release of ATP [97]. In addition to this, HIF-1α in astrocytes is reported to be upregulated dependent on ATP binding to P2X7R in response to mild ischemia [98]. Acting as the sensor of ischemic tolerance, P2X7R provides neuroprotection against subsequent severe ischemia [99]. These events are all associated with ATP/P2X7R/ HIF-1α signals in astrocytic activation, suggesting the importance of purinergic signaling for neuroprotection. For further understanding metabolic reprogramming mediated by purinergic signaling in the nervous system, the study of downstream pathways in the brain is urgently needed.

The more we learn about purinergic receptors, the more we realize that purinergic signaling extends well beyond the central or peripheral nervous systems, and in fact, its participation in inflammation and immune response might be even more relevant and pervasive. In these scenarios, P2X7R stands out as a leading player and a promising target for innovative therapeutics [53].

ATP plays a widely accepted role in self-defense, among plants and mammals, as one of the most ancient and conserved danger-associated molecular patterns (DAMPs). DAMPs include a unique class of molecules that are normally sequestered intracellularly and are released only upon cell damage. Once in the extracellular space, DAMPs elicit danger signals and activate the immune system by interacting with specific receptors. P2X7R is unanimously recognized as the main sensor for ATP during inflammation and the main trigger of pro-inflammatory effects [100–102]. Accordingly, its inhibitor, oxidized ATP, is an efficient immune suppressor for preventing transplant rejection [103, 104]. Following a pro-inflammatory response, activation of the inflammasome, a multiprotein complex, is required for an immune response [105]. In addition, negative regulation of GSK3β activity is an accepted mechanism by which P2X7R might contribute to T cell proliferation downstream of T lymphocyte activation [106, 107], since the expression of constitutively active GSK3β decreases T cell proliferation, differentiation, and survival [108, 109]. However, GSK3β is also reported to mediate metabolic reprogramming, specifically enhancing glycolysis and pyruvate metabolism, in TGF-β1-treated ARPE-19 cells [110]. These contradictory findings put us in a quandary such that we cannot exactly determine how GSK works in ATP/P2X7R-mediated metabolic reprogramming. On the other hand, P2X7R is linked to HIF-1α by a biunivocal relationship in which active P2X7R stimulates HIF-1α, but in turn, HIF-1α activation upregulates P2X7R expression [111]. Finally, it remains unclear whether the generation of immunological memory is impacted by the same pathway, but evidence has shown that P2X7R, indeed, plays a hitherto unsuspected and intrinsic role in supporting the generation of long-lived memory CD8⁺ T cells by driving their mitochondrial maintenance and metabolic reprogramming, in part, in an AMPK-activated manner [112].

The proliferative effect of P2X7R upregulation is described in human leukemic lymphocytes and human neuroblastoma cells [113–115], which relates to an enhanced mitochondrial energy state [116, 117]. P2X7R acts as a key determinant of tumor microenvironment composition due to its combined action on ectonucleotidases, ATP release, and immune cell infiltration [118, 119]. These findings characterize P2X7R as a promising therapeutic target in oncology. Small drug-like P2X7R antagonists have already been taken to clinical trial for select chronic inflammatory diseases. However, rather disappointing results regarding clinical efficacy were obtained [120–122]. The unsatisfactory outcomes may prompt the use of P2X7R-targeted biologics. Anti-P2X7R antibodies are currently under development [123, 124]. Better P2X7R modulators than small-drug agonists may be found.

Interestingly, recent research has demonstrated that ATP release from chemotherapy-treated tumor cells is a potent driver of the antitumor immune response through P2X7R activation on dendritic cells (DCs), which are induced to complete maturation and acquire full competence during antigen presentation [102, 125, 126]. In addition, chemotherapy treatment for AML may result in the induction of an immune suppressive microenvironment. This scenario, at least in part, depends on the local effects of ATP release from dying leukemia cells. And the released ATP may lead to an increase of tolerogenic DCs and leukemia-associated Tregs in a P2X7R-dependent manner [127]. ATP may directly elicit the Treg suppressive function or may be metabolized into adenosine, which enhances Treg activity [119, 128].

Arguably, P2X7R plays an essential role in the setting of chronic stress, which is known to contribute to depression in humans due to the priming of inflammation. In rodents exposed to chronic unpredictable stressors, increases in extracellular ATP release in the hippocampus were observed [129]. Furthermore, antagonistic action and genetic knockout of P2X7 prevented the development of depression-like behaviors in this model [129]. And a similar role for P2X7 has been observed in schizophrenia [130]. All the facts indicate that ATP and P2X7R can, again, be expected to have involvement in the development of this disease, given the strong evidence that purinergic signaling is increasingly involved during the pathology of psychiatric disorders [131, 132]. However, whether metabolic reprogramming participates in psychiatric conditions remains unclear.

In addition to the controversial interaction between biological processes and stimulated ATP-P2X7R signals, further studies reveal how P2X7R works to initiate and regulate metabolic reprogramming. The expression of P2X7R was found to enable better adaptability to unfavorable ambient conditions via upregulation of glycolysis (P2X7R-expressing cells overexpressing PFK and PDHK1) and intracellular glycogen stores (HEK293-P2X7 cells induce the expression of the ubiquitous glucose transporters Glut1 and G3PDH). P2X7R confers a substantial growth advantage by activating a series of important intracellular pathways that contribute to the generation of the energy and components required for the synthesis of amino acids, lipids, and nucleic acids [117].

As for P2YRs and metabolic reprogramming, Yvan first reported that P2YRs are involved in the inhibitory action of ATP on glucose transport [133]. More recently, P2Y6R activation by UDP or MRS2957 was also shown to increase GLUT4 translocation and glucose uptake in primary 3T3-L1 cells [134]. Under pathological conditions, P2YRs were reported to play a regulatory role in metabolic reprogramming in type 2 diabetes mellitus (T2DM) development. In pancreatic islets of humans and other species, P2YRs’ activation has been shown to stimulate insulin secretion in the glucose-dependent manner [135, 136].

An overview of current evidence about the regulatory effects of extracellular purines on metabolic reprogramming is shown in Table 2.

Table. 2.

Extracellular purines and metabolic reprogramming

Receptors	Cell types	Biological process	Metabolic alterations	Reference
P1 receptors
ADORA2A ADORA2B	Macrophage	Activation and inflammatory response	Glycolytic flux, glycolysis, and OXPHOS	[22, 71, 72]
Unknown	DCs, tumor-associated macrophages	Immunosuppressive function, pro-tumor effects	Glycolysis and OXPHOS	[38, 74]
ADORA2A	Brown adipocytes	Brown adipose tissue activation, thermogenesis	Lipolysis	[73]
ADORA2B	Erythrocytes	Oxygen release	Glycolytic flux and glycolysis	[30, 31]
ADORA2B	Hematopoietic progenitor cells	Stress hematopoiesis	Glycolytic flux and glycolysis	[77]
P2 receptors
P2X7R	HEK293 cells	Cell growth	Glycolysis and glycogen stores	[92]
P2YRs	WA and 3T3-L1 cells	Unknown	Glycolytic flux and metabolism	[93, 94]

OXPHOS, oxidative phosphorylation

Intracellular pathways

AMP pathways

One of the fundamental requirements of all cells is the ability to balance ATP generation and consumption. AMPK, regulated by the [AMP]/[ATP] ratio, is well established as a metabolic regulator [64, 65, 137]. As the [AMP]/[ATP] ratio increases, AMP binds to AMPKγ, which subsequently leads to conformational changes that enable phosphorylation of the α subunit by kinases upstream of AMPK, liver kinase B 1 (LKB1), and calcium/calmodulin-dependent protein kinase kinase-β (CaMKKβ), to activate AMPK [138, 139]. Falling energy levels lead to AMPK activation in an AMP-dependent manner, favoring ATP production by stimulating fatty acid oxidation (FAO), glucose uptake, and glycolysis, as well as by reducing ATP-consuming processes such as transcription and protein synthesis, depending on the tissues and the triggers [140].

In skeletal muscle during prolonged exercise, the AMP-AMPK pathway, which is necessary for efficient muscle contraction, promotes mitochondrial biogenesis and nutrient uptake [65]. In addition to influencing metabolite transport, the pathway also promotes FAO via phosphorylation and inactivation of the second isoform of acetyl-CoA carboxylase (ACC2) [65], leading to decreased malonyl-CoA generation and increased transport of fatty acids. This mechanism has been suggested to exist in multiple cell types, including adipocytes, macrophages, and hepatocytes [141, 142].

AMPK promotes systemic and cellular immunometabolism in an organism. Researchers demonstrated that AMPKα1 (an overriding AMPK isoform in T cells) is essential for Th1 and Th17 cell proliferation and primary T cell responses to viral and bacterial infections in vivo, indicating that AMPK-dependent regulation of metabolic homeostasis is a key modulator of T cell–mediated adaptive immunity [143]. Treatment with (aminooxy) acetic acid (AOA) may increase the phosphorylation of AMPKα in Th17 cells without affecting the mammalian/mechanistic target of rapamycin complex 1 (mTOR) activity [144]. In contrast, pentanoate, a physiologically abundant short-chain fatty acid, likely enhances glycolytic metabolism via the inhibition of AMPK in Th17 cells [145]. All these facts uncovered the complicated underlying mechanism of AMPK-modulating metabolism alterations of T helper cells and have contributed to proposed new perspectives and challenges. A study of AMPK-deficient CD8⁺ T cells showed that the generation of memory T cells is dependent on AMPK-induced upregulation of oxidative phosphorylation, mitochondrial biogenesis, and FAO [146, 147]. R. D. Michalek used in vitro analysis and showed that AMPK is stimulated in murine TCR-activated Tregs, thereby inhibiting mTOR and promoting lipid oxidation instead of glycolysis. Activated AMPK and targeted induction of lipid metabolism were shown to enhance the number of Tregs in vivo [148]. Collectively, AMPK is recognized as a checkpoint for activated T cells, regulating mRNA translation and glutamine-dependent mitochondrial metabolism to maintain T cell bioenergetics and viability [149]. However, loss of AMPK was found to cause only a slight disturbance in T cell homeostasis in LKB1−/−T cells [143, 150], suggesting that AMPK-dependent regulation of metabolic homeostasis is indeed a modulator of T cell–mediated adaptive immunity, but its specific role remains controversial. Similar to that of T cells, B cell metabolism is also reported to be linked to AMPK activation, which prevents the differentiation of unswitched memory B cells to plasmablasts but supports their differentiation into CD27−IgD− B cells by inhibiting mTOR [23]. Depletion of glucose can also appropriately limit macrophage and effector T cell–driven inflammation and concomitantly initiate M2 macrophage and Treg-dependent TGF-β secretion upon activation of AMPK signaling [151, 152].

High glycolytic flux, coupled with a decreased vascular supply, leads to profound glucose depletion within tumors [153]. Glucose depletion has a profound impact on the function of surrounding immune cells by favoring the growth and differentiation of Tregs, which have higher AMPK activity, lower glucose oxidation, and greater fatty acid oxidation to support energy homeostasis [151, 154, 155]. AMPK is found to have an effect on ovarian cancer [156]. Impaired mitochondrial respiration resulting from glucose deprivation and limited influx of glutamine is also involved in tumor-infiltrating T cells, making them dysfunctional and incapable of controlling malignant progression [157]. Stromal fibroblasts, which can sustain the viability of cancer cells and enhance malignant transformation, can also be reprogrammed by glucose depletion to support tumor growth [158]. Glucose depletion and subsequent AMPK activation in fibroblasts might activate autophagy, driving the secretion of nonessential amino acids (NEAAs) that are significantly depleted in tumor microenvironment [159, 160]. All these facts indicate that AMPK behalves a central role in the metabolic reprogramming of tumor.

Tightly linked to inflammation, tumor, and other pathological states, hypoxia induces ATP depletion and reactive oxygen species (ROS) accumulation [161], which directly activates AMPK and consequently mediates metabolic rewiring to promote survival during energy restriction through activation of catabolic pathways (glucose uptake, glycolysis, FAO, and gluconeogenesis) and suppression of anabolic pathways (protein, glycogen, and fatty acid synthesis) [65, 137, 162, 163]. AMPK is also responsible for the regulation of the redox state by alleviating glucose deprivation–induced NADPH deficiency via decreased FAS and increased FAO [137]. Of note, researchers thought it reasonable to propose that hypoxia could collaboratively drive cancer-related fibroblast function with glucose deprivation [158].

Recent studies revealed that the decline in the sensitivity and responsiveness of AMPK due to age had been shown to be a factor in many age-associated diseases, including cardiovascular diseases, metabolic syndrome, and even Alzheimer disease [164]. However, whether metabolic reprogramming has a role in these diseases is a subject of debate.

Therefore, we wondered how AMPK achieves its metabolism-altering effects. Phosphorylated AMPK could lead to the induction of cellular NAD+ levels, resulting in activation of sirtuin 1 (SIRT1)—an NAD+-dependent deacetylase that positively regulates PGC-1α such that mitochondrial energy metabolism and biogenesis are activated [139, 165, 166]. It is highly suspected that the regulatory role of AMPK is linked to downstream proteins such as mTOR, mitochondrial uncoupling protein 2 (UCP2), carnitine palmitoyltransferase 1C (CPT1C), and p53 [167, 168].

Studies have suggested that treatment with rapamycin, which inhibits mTOR and glycolysis, may be a useful strategy to restrict tumor growth by reversing metabolic reprogramming in tumor cells. However, studies have shown that untargeted rapamycin treatment also impairs the ability of tumor-infiltrating NK cells to kill tumor cells, similar to the effect of obesity [35].

As a highly expressed mitochondrial membrane protein in a variety of tumor cell types, UCP2 is capable of uncoupling mitochondrial respiration and protecting against ROS by reducing mitochondrial respiration [169, 170]. It also serves as a key regulator of energy metabolism that limits glucose oxidation while promoting the oxidation of alternative substrates such as glutamine and fatty acids [171–174]. Activation of AMPK has been shown to be involved in the upregulation of UCP2 expression [175, 176], whereas UCP2 overexpression has been shown to increase signaling from AMPK [177], although in both cases, the underlying mechanisms remain unclear. Based on established facts, activation of AMPK signaling in UCP2-overexpressing cells is associated with a downregulation of HIF expression [177], and hypoxia decreased UCP2 expression via HIF-1-mediated suppression of PPARγ [178]. A recent study clearly indicated that defects in MTO1, the mitochondrial tRNA modification enzyme, lead to reprogramming of cell metabolism mediated by the HIF-PPARγ-UCP2-AMPK axis. MTO1 fibroblasts exhibit HIF-1 activation, downregulation of PPARγ, UCP2, and inactivation of AMPK. Due to metabolic reprogramming, the utilization of fatty acids for the de novo synthesis of fatty acids and β-oxidation appears to be compromised, leading to the accumulation of lipid droplets in MTO1 fibroblasts. However, cells depleted of the GTPBP3 protein produce the opposite effect [179]. In addition, UCP2 has also been found to accommodate a metabolic shift from glucose oxidation to FAO and induce ketogenesis in an AMPK-related manner, resulting in neuron viability alteration and progressive neurodegeneration [180].

Theoretically, competing for AMPK binding sites via the use of an agonist could lead to metabolic turbulence in tumor cells. Metformin, the most widely prescribed T2DM drug, has been shown to modulate metabolism by acting as a mild inhibitor of Complex I in the respiratory chain [181]. Recently, metformin was recognized for its ability to induce ablation of miR-21-5p, leading to the activation of AMPK in addition to the inhibition of mTOR [182], thus showing therapeutic potential for breast cancer. However, Cristina Oliveras-Ferraros and colleagues established a preclinical model of estrogen-dependent MCF-7 cells that were adapted to grow chronically (> 10 months) in the presence of a graded, millimolar mixture of metformin. They found that breast cancer cells, in which metabolic reprogramming has been verified to be related to AMPK, showed acquired resistance to metformin [183, 184]. In other critical findings, metformin could suppress downstream signals such as those for the hypoxia-induced stabilization of HIF-1 by reprogramming oxygen metabolism in hepatocellular carcinoma and playing an AMPK-independent role in T cells [185, 186]. All the facts suggest that metformin may miss the target during clinical treatment by promoting acquired resistance and creating side effects. As another AMPK agonist, 5-aminimidazole-4-carboxamide1-β-D-ribofuranoside (AICAR) is a cell-permeable precursor to ZMP that simulates AMP to bind to the AMPKγ subunit [187]. Finally, a number of natural compounds, such as resveratrol and curcumin, reportedly activate AMPK and confer similar beneficial effects on metabolic reprogramming-related diseases, such as metformin AICAR [185, 188, 189], creating a bright future for the treatment of metabolic diseases.

An illustration of the AMP/AMPK pathway for regulating metabolic reprogramming is shown in Fig. 2.

Fig. 2.

The regulatory role AMP plays in metabolic reprograming via activation of AMPK: under hypoxia, glucose deprivation and prolonged exercise in multiple tissues, the [AMP]/[ATP] ratio is elevated, thus initiating direct AMP binding to AMPKγ and stimulating AMPK. The activation of AMPK mediates metabolic reprogramming to promote cell survival during energy restriction by promoting catabolic pathways and suppressing anabolic pathways. Specifically, activated AMPK leads to the stimulation of SIRT1 which positively regulates PGC-1α and subsequent mitochondrial energy metabolism and biogenesis. AMPK activity can sustain glycolysis via the direct inhibition of mTOR. AMPK activation also promotes FAO via phosphorylation and inactivation of ACC2, which limits malonyl-CoA generation. Hypoxia elicits metabolic reprogramming via the accumulation of ROS or the HIF-PPARγ-UCP2-AMPK axis. UCP2 overexpression has been shown to increase signaling from AMPK, resulting in the limitation of glucose oxidation and promotion of FAO and ketogenesis. The AMPK agonist metformin activates AMPK by increasing [AMP]/[ATP] ratio. Another agonist, AICAR, simulates AMP to bind to AMPKγ

Intracellular cyclic purines

Regarding metabolism, the cAMP/PKA pathway has been confirmed to promote metabolic reprogramming in different cell types [190, 191], even in those of nonmammals such as Trypanosoma brucei [192]. Through metabolomics, proteomics, and other omics, Cheng and coworkers proposed that the metabolic reprogramming mediated by cAMP/PKA contributes to the adaptation of Escherichia coli to glycerol [193]. For mammals, the cAMP/PKA pathway enhances the uptake of glucose for metabolic rewiring, thus assisting in the prevention of hepatic steatosis [194, 195]. Additionally, cAMP-mediated metabolic effects may be closely linked to the impact of agmatine during calorie restriction [196]. Lu and coworkers managed to increase glucose uptake by upregulating GLUT expression in corticotropinoma after exposure of AtT-20 cells to 8-bromoadenosine 30,50-cyclic monophosphate (8-Br-cAMP, an analogue of cAMP) [194].

Some transcription factors might be responsible for cAMP-regulated metabolic reprogramming. For example, as an enzyme downstream from AMPK, SIRT1 could also be phosphorylated by PKA and thus induce modification during lipolysis [190]. It is considered a potential mechanism in the physiological thermogenic process. Furthermore, it is reported that cAMP-mediated HIF-1α activation could enhance the glycolysis of pancreatic β-cells and upregulate insulin secretion, creating a promising future for diabetes therapy treatments [197]. However, unlike the potential role of regulatory cAMP in metabolic reprogramming [71], little is known about whether cGMP participates in regulating metabolic reprogramming.

Although a seemingly similar mechanism, cyclic purines function intracellularly, and the various upstream pathways make their effects various and complicated. Together with downstream tissue-specific effector molecules and cellular heterogeneity, targeting cyclic purines might not be sufficient or efficient for metabolic reprogramming manipulation.

Conclusion

In summary, apart from the maintenance of homeostasis, purinergic signaling certainly plays a remarkable regulatory role in metabolic reprogramming. It is worth noting that purines could also be generated within metabolic reprogramming, forming a positive purine-metabolite reprogramming feedback pathway. Intriguingly, although both purinergic signaling and metabolic reprogramming may be ubiquitous, current studies of their relationships are focusing on some classic topics such as oncology or immunology. Therefore, more investigations about other popular themes such as autophagy and aging-related diseases might be required to expand the discussion about these relationships.

On the other hand, how underlying mechanisms express identical purinergic signaling function differently in different cellular metabolism pathways remains a perplexing issue. Due to heterogeneity, purine mediates metabolic reprogramming in various ways. Therefore, to clarify the specific pathways in certain cells and provide targets for diagnosis and treatment, new single-cell metabolomics analysis techniques are in urgent demand. Even though plenty of mysteries about the impact of purines on metabolic reprogramming have made the underlying mechanism complicated, advancements have revealed more surprises in solving some confusion with regard to immunotherapeutic resistance and hematopoiesis regulation. Because of the prosperity from and progress in drugs targeting purinergic signaling, metabolic reprogramming or both, we see a bright future for the translation of these advances.

A summary of the regulatory network of purines, metabolic reprogramming, and biological processes is illustrated in Fig. 3.

Fig. 3.

A summary of the regulatory network of purines, metabolic reprogramming, and biological process: the dark spots above represent the purines and the light spots represent the various biological process. Finally, the lines linking these two balls represent the metabolic pathways that are regulated by purines and result in the functional alteration. The types of metabolic pathways are represented by the colors of lines, which are illustrated in the lower right corner

Funding information

This work was financially supported by the grants from the Health and Family Planning Commission of Hunan Province (B20180855), Innovation Driven Planning of Central South University(2018CX028), and High-level Talent Planning of Xiangya Hospital.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Human participants and/or animals

This article does not contain any studies with human participants performed by any of the authors.

Informed consent

Informed consent was obtained from all individual participants included in the study.