The hypoxia-adenosine link during inflammation

Abstract

Hypoxic tissue conditions occur during a number of inflammatory diseases and are associated with the breakdown of barriers and induction of proinflammatory responses. At the same time, hypoxia is also known to induce several adaptive and tissue-protective pathways that dampen inflammation and protect tissue integrity. Hypoxia-inducible factors (HIFs) that are stabilized during inflammatory or hypoxic conditions are at the center of mediating these responses. In the past decade, several genes regulating extracellular adenosine metabolism and signaling have been identified as being direct targets of HIFs. Here, we discuss the relationship between inflammation, hypoxia, and adenosine and that HIF-driven adenosine metabolism and signaling is essential in providing tissue protection during inflammatory conditions, including myocardial injury, inflammatory bowel disease, and acute lung injury. We also discuss how the hypoxia-adenosine link can be targeted therapeutically in patients as a future treatment approach for inflammatory diseases.

hypoxia-inducible factors (HIFs) are found stabilized during a number of inflammatory conditions and diseases, including inflammatory bowel disease, pathogen infection, acute lung injury, myocardial injury, or during ischemia-reperfusion injury (45). What contributes to tissue hypoxia during inflammation? On the one hand, inflammation causes alterations in tissue metabolism and is characterized by a profound increase in local oxygen demand. On the other, ischemia, damage to tissue blood supply (trauma), thrombosis, or compression (interstitial hypertension) reduces the supply of oxygen and other metabolic substrates (44). Inflammatory bowel disease (IBD) serves as an elegant example of inflammation-associated hypoxia (Fig. 1) (45). In the gut, intestinal epithelial cells are at baseline conditions already hypoxic. However, with the onset of inflammation, as occurs with IBD, gut tissues become profoundly hypoxic (71). Organ transplant is a primary example of hypoxia-associated inflammation. The degree of ischemic injury in donor kidneys (78) or transplanted lungs (5, 28) is shown to correlate with increased expression of toll-like receptor (TLR) 4, an inflammation initiating receptor, and increased inflammatory cytokines; increased TLR4 and inflammatory cytokines correlates with loss of graft function as well. Relevant to this review, studies in adenosine metabolism/signaling gene-deficient mice show that short-term, low oxygen exposure increases vascular leakage, proinflammatory cytokine expression, and the accumulation of inflammatory cells in tissues (33, 43, 46, 119, 135). Suffice to say, hypoxia and inflammation are closely linked: on the cellular level, hypoxia can cause inflammation, and inflammation can cause hypoxia. Important to emphasize is that although hypoxia can be an inflammatory stimulus that encourages proinflammatory responses and breaks down tissue barriers, there are many examples in which stabilization of HIFs induces anti-inflammatory and tissue-protective responses. Indeed, extracellular adenosine metabolism and signaling is an essential HIF-driven pathway that dampens excessive inflammatory responses and induces barrier protection in conditions of low oxygen and inflammation. Of note, not all situations of hypoxia are associated with a primary role of inflammation. High-altitude pulmonary edema is the resultant of hypoxia pulmonary vasoconstriction and occurs in the absence of inflammation; inflammation follows the injury to the alveolar capillaries (7, 90, 126, 127). In this review, we discuss the relationship between inflammation, hypoxia, and adenosine and that HIF-driven adenosine metabolism and signaling is essential in providing tissue protection during inflammatory conditions, including myocardial injury, inflammatory bowel disease, and acute lung injury (Fig. 1). We also discuss how the hypoxia-adenosine link can be targeted therapeutically in patients as a future treatment approach for inflammatory diseases.

Fig. 1.

Adenosine metabolism and signaling in human health conditions and diseases characterized by hypoxia-associated inflammation or inflammation-associated hypoxia. Right: hypoxia-driven adenosine metabolism and signaling in intestinal inflammation. Inflammatory bowel disease (IBD) is characterized by excessive inflammation that severely damages the mucosa. Inflammation is associated with a severe shift in oxygen supply and demand, resulting in profound hypoxia. With inflammation and hypoxia, multiple cells release ATP/ADP, including inflammatory cells, platelets, and epithelial cells. Activation of the ATP/ADP receptor P2X7 on enteric neuron cells further promotes inflammation and tissue injury [not shown (58)]. Hypoxia results in an Sp1-dependent induction of CD39 [not shown (47)] and a hypoxia-inducible factor (HIF)-dependent induction of CD73 (128) and A2BAR (76). CD39 is largely restricted to immune cells and vascular endothelium. Epithelial cell expression of ectonucleoside triphosphate diphosphohydrolase (NTPDase) activity may be related to CD39 family member ENTPD7 (83) and alkaline phosphatases (97). Together, increased CD39, CD73, and A2BAR support increased adenosine metabolism and signaling. Adenosine signaling, particularly involving A2AAR (101) and A2BAR (53), dampens inflammatory responses, such as tissue infiltration of neutrophils, and promotes protection of the barrier. [Modified from Eltzschig et al. (48) with permission. Copyright 2012 Massachusetts Medical Society.]

Adenosine Signaling During Inflammation

Inflammation is devastating to tissues. Mounting an inflammatory response, however, is essential for eliminating harmful stimuli, such as pathogens or irritants, clearing out damaged cells from the initial cause of injury, and to initiate tissue repair. Equally essential are responses that exist to dampen potentially harmful, unrestricted inflammatory responses, minimizing contralateral damage to tissues and promoting tissue integrity. Therefore, once inflammation is initiated, bringing to an end the inflammatory response becomes as important as eliminating the inflammatory pathogen or insult. Adenosine signaling has emerged as being crucial to dampening immune responses. Nearly a century ago, observations by Drury and Szent-Gyorgyi (31) of an adenine compound from cardiac extracts to slow heart rate provided the first indications of extracellular adenosine as a signaling molecule. Adenosine belongs to a group of purinergic signaling molecules that also include ATP and ADP (42). Once released to the extracellular surface, ATP/ADP is rapidly hydrolyzed to adenosine by cell surface enzymes (42). Ecto-nucleoside triphosphate diphosphohydrolases (E-NTPDases), including E-NTPDase 1 (CD39), metabolize ATP/ADP to AMP, and ecto-5′-nucleotidase (CD73) generates extracellular adenosine from AMP. Adenosine can signal through one of four membrane-spanning adenosine receptors: A1AR, A2AAR, A2BAR, and A3AR, or re-enter the cell through equilibrative or concentrative nucleoside transporters (ENTs or CNTs, respectively). The movement of extracellular adenosine into the cell though ENTs or CNTs acts to dampen adenosine signaling (42, 56). Extracellular adenosine may also be deaminated by cell surface CD26-conjugated adenosine deaminase (69) as well as phosphorylated back to a nucleotide; both also serve in terminating adenosine signaling.

In the past several decades, pharmacological and genetic studies have contributed significantly to understanding adenosine signaling during inflammation. Two major concepts emerged: 1) adenosine signaling is central to dampening inflammatory responses and subsequently protects tissues; and 2) adenosine receptors, especially A2AAR and A2BAR, are involved in limiting immune responses. For example, in vivo studies from Cronstein et al. (20) showed the injection of mice with methotrexate raised extracellular adenosine levels and limited immune cell accumulation in inflamed tissues. The reduction in immune cell accumulation was completely reversed by an A2AAR antagonist. Studies in A2AAR^−/− mice from the laboratory of Sitkovsky (105) confirmed these findings utilizing in vivo models of inflammation. It is now well recognized that inflammation and conditions of low oxygen increase extracellular ATP/ADP release, which substantially raises extracellular adenosine levels (48, 66). As discussed later in this review, raising extracellular adenosine levels is supported even more by transcriptional regulation of adenosine metabolizing and receptor genes via hypoxia-inducible factors (HIFs) (3, 43, 76, 98, 99, 128). The importance of extracellular adenosine metabolism is seen in CD39- and CD73-deficient mice, which show exaggerated inflammation, and in that inflammation is significantly suppressed in mice exposed to adenosine signaling terminating pathway member (e.g., ENTs, AK) inhibitors (43, 98, 99). Adenosine signaling via A2AAR has shown beneficial anti-inflammatory effects in models of ischemic or traumatic tissue damage in various tissues, including liver (27), kidney (26), intestine (101), heart (86, 141), skin (109), and lung (114).

In general, extracellular adenosine dampening inflammation via A2AAR involves its activation on immune cells. For example, A2AAR activity on macrophages limits their production of proinflammatory cytokines, such as interleukin-12 (IL-12) and tumor necrosis factor-α (TNF-α) (62, 77) and promotes their release of anti-inflammatory cytokine, interleukin-10 (IL-10) (102). Similarly, A2AAR activity on mature dendritic cells shifts their cytokine profile from proinflammatory to anti-inflammatory, reducing IL-12, interleukin-6 (IL-6), and interferon-α (IFN-α) production and increasing IL-10 production (107, 108, 122). A number of studies have defined that A2AAR activity on neutrophils (en route to sites of inflammation) prevents tissue injury by inhibiting the production of toxic oxygen species (18, 19, 21, 137) as well as modulates the production and release of proinflammatory cytokines (e.g., TNF-α), chemokines (94, 133), and prostaglandins (e.g., PGE₂) (12, 111), and inhibits production of chemoattractant leukotrienes (e.g., LTB₄) (79, 80). Considering T cells, activation of A2AAR on effector T cells inhibits their activation, differentiation, and effector functions, whereas A2AAR is implicated in enhancing the formation of T regulatory (T_reg) cells and their expression of program cell death protein 1 (PD-1) and CTLA-4, both negative regulators of inflammation (104). Studies by Robson and colleagues (29) have shown an important regulatory loop between T_regs and effector T cells in that CD39/CD73 coordinated adenosine metabolism on T_regs activates A2AAR on effector T cells, dampening effector T cell activity. A2AAR activity on vascular endothelial cells has important anti-inflammatory effects. A2AAR activation inhibits endothelial cell release of inflammatory cytokines and expression of adhesion molecules, such as E-selectin and vascular cell adhesion molecule 1 (VCAM-1) (9), whereas A2BAR activity increases the vascular barrier (33, 87). Collectively, this limits the accumulation of neutrophils into tissues and subsequent tissue damage.

A2BAR also has important anti-inflammatory actions on immune cells. Similar to A2AAR, A2BAR activity on macrophages limits their production of inflammatory cytokines, such as TNF-α (77). A2BAR is also significant in protecting the epithelial barrier during inflammation (2, 60). The anti-inflammatory signaling molecule netrin-1’s engagement of A2BAR also dampens infiltration of neutrophils in models of lipopolysaccharide (LPS)-induced acute lung injury (96), hypoxia-induced lung inflammation (119), and acute experimental colitis (1). Taken together, extracellular adenosine metabolism and signaling is crucial in dampening inflammatory responses and involves shifting immune responses from proinflammatory to anti-inflammatory as well as protecting tissue integrity through inducing barrier protection.

Mechanisms of Hypoxia Adaptation

When oxygen becomes limited in tissues, such as occurs with inflammation, there is a need to be able to detect hypoxia as well as to initiate responses to adapt to the conditions. Recently, William Kaelin, Peter Ratcliffe, and Gregg Semenza were awarded the 2016 Albert Lasker Basic Medical Research Award for their exceptional discoveries that pioneered and has deepened our understanding of the fundamental pathways by which cells sense and respond to the presence of low oxygen. HIFs are αβ-heterodimeric transcription factors that are essential to hypoxia adaptation, having a central importance in orchestrating transcriptional programs that regulate tissue metabolism and maintenance of homeostasis in conditions of low oxygen.

When oxygen levels are sufficient in tissues, hypoxia-inducible factor-1 (HIF-1)-α and HIF-2α experience posttranscriptional iron- and oxygen-dependent hydroxylation, which occurs on two proline residues within the oxygen-dependent degradation (ODD) domain of the α-subunit. The α-subunit is recognized by the von Hippel-Lindau (VHL) tumor suppressor protein (pVHL) E3 ubiquitin ligase complex and sent for proteasome degradation (50, 68, 70, 103, 130). In hypoxic conditions, the hydroxylation of HIF-α subunits does not occur, which stabilizes HIF1α and HIF2α. With stabilization, HIF1α and HIF2α translocate to the nucleus and form a complex with the β-subunit and other coactivations and begin transcriptional regulation by binding to hypoxia-responsive elements (HRE) of target genes (123, 124). Over 2% of all human genes are direct target genes of HIFs and include genes that encode glycolytic enzymes, inhibit mitochondrial respiration, regulate apoptosis, modulate inflammation, and promote angiogenesis (92). Although HIF-1α and HIF-2α are closely related, knockout studies in mice have shown them to have nonredundant roles and distinct phenotypes (113). Tissue specificity and temporal expression of the isoforms are considered to account for the differences (64, 113, 118, 139). However, distinct transcriptional gene targets are more likely the source of the difference (65, 113). For example, the transcriptional gene regulation of glycolytic pathway enzymes appears to be driven by HIF-1α as opposed to HIF-2α (65). More so, the ability of HIF-2α to be involved in an oxygen-regulated translation initiation complex has been described (136). Hypoxia stimulates the formation of a complex involving HIF-2α, RNA-binding protein RBM4, and the cap-binding eIF4E2. The complex captures mRNAs and targets them for active translation, thus evading hypoxia-induced repression of protein synthesis. HIF-3α response in hypoxia is an area of active study (142).

The physical interaction of HIF-α subunits and pVHL is importantly controlled by prolyl hydroxylation. Prolyl hydroxylases (PHDs) were discovered to be essential molecular oxygen sensors after years of searching for the mechanism through which hypoxia stabilizes HIFs (50, 67, 68). In conditions of low oxygen, PHDs are inactive or have reduced activity since they require oxygen as a cofactor, which provides a clear connection between the availability of oxygen and the regulation of HIFs. In mammalian cells, three isoforms of PHD exists (PHD1, PHD2, and PHD3), which hydroxylate HIF-α and target HIF-α for proteasome degradation involving pVHL (50, 67, 68). Another level of control lies with asparaginyl hydroxylase factor inhibiting HIF (FIH), which is also oxygen-sensitive and iron dependent (84). Low oxygen levels block the hydroxylation of a conserved asparagine residue, facilitating the recruitment of transcriptional cofactors, such as p300, that are necessary for HIF-mediated gene transcription (85). Accordingly, both PHDs and FIH serve important roles as oxygen sensors by sensing the presence or absence of oxygen and controlling the activity of HIFs. HIF-driven gene programs have been shown to induce anti-inflammatory responses in a number of inflammatory disease models, for example, through the increased production and signaling effects of anti-inflammatory molecules, such as adenosine (3, 43, 76, 98, 99, 128) and netrin-1 (119).

Hypoxia Signaling During Inflammation

Inflammatory conditions are often characterized by hypoxia, whereas conditions of low oxygen are often characterized by the onset of tissue inflammation. Inflammation-associated hypoxia or hypoxia-associated inflammation is implicated in a number of human health conditions and diseases, including cancer, organ transplantation, and obesity as well as acute lung injury, IBD, and pathogen infection (45). Indeed, gut tissues of murine models of IBD, a disease characterized by excessive inflammation, are found to be profoundly hypoxic and show HIF stabilization (16, 71). In models of sepsis, proinflammatory factors, such as lipopolysaccharide (LPS), have been shown to directly stabilize HIFs (110). Here, LPS increases HIF-1α stability and decreases PHD in a Toll-like receptor (TLR)-dependent manner, specifically involving TLR4. As well, HIFs can also induce TLR expression, such as TLR2 and TLR6 (81). Several studies show evidence for HIF-1α stabilization during bacterial infections. For example, bacterial release of iron-binding siderophores, as such with Yersinia enterocolitica, can stabilize HIF by inhibiting PHD through forming chelate complexes with iron ions (61). Additionally, the citric acid cycle intermediate, succinate, can stabilize HIF-1α during LPS-induced inflammation (131).

Other studies show that NF-κB (nuclear factor κ-light-chain-enhancer of activated B cells), a transcription factor important to inducing proinflammatory responses to pathogens, is intimately linked to hypoxia signaling. Studies by Taylor (22) suggested that PHDs may link oxygen sensing to NF-κB by IKK (IκB kinase complex). These studies identified that IKKβ (subunit of IκB kinase complex) contains a conserved prolyl hydroxylation site similar to HIF-1α (22). Hypoxia increases IKKβ stability through inhibition of IKKβ hydroxylation by PHD1, which directs NF-κB to disassociate from its inhibitor protein, thus promoting NF-κB to induce proinflammatory gene expression. Importantly, these studies demonstrated that oxygen-dependent PHDs are not restricted to regulating HIF-α to direct gene programs during inflammation and hypoxia. Additionally, similar to PHDs, the asparaginyl hydroxylase FIH has also been found to efficiently hydroxylate proteins of the IκB kinase family (15). Many more studies show that inflammatory stimuli can induce HIF-1α expression in a NF-κB-dependent manner, whereas, in turn, NF-κB activity in hypoxia can be regulated by HIF-1α (25).

Studies involving Type 1 regulatory cells (Tr1) and aryl hydrocarbon receptor (AHR) have provided additional insight into hypoxia signaling in modulating immune responses in inflamed tissues. Mascanfroni, Quintana, and colleagues (93) recently reported HIF-1α controls the early metabolic reprograming of Tr1 cells, whereas Tr1 cell metabolism is later controlled by AHR. AHR promotes HIF-1α degradation and increases CD39 expression. Extracellular ATP and hypoxia, both linked to inflammation, trigger AHR inactivation by HIF-1α and inhibit Tr1 cell differentiation. Conversely, CD39 promotes Tr1 cell differentiation by depleting extracellular ATP. Here, HIF-1β and HIF-1α interactions increase and the binding of AHR to HIF-1β decreases. AHR stability is reduced by being targeted to the proteasome, via HIF-1α, which decreases the recruitment of AHR to the CD39 promoter. Shift in control, i.e., HIF-1α to AHR, occurs with the upregulation of PHD proteins, promoted by AHR, which encourages HIF-1α degradation and the subsequent rise in CD39 expression and extracellular ATP depletion. Altogether, these data suggest that HIF-1α and AHR function to integrate immunological, metabolic, and environmental signals to regulate Tr1-type immune responses (93). Taken together, these findings support that hypoxia and inflammation share a close relationship. Although hypoxia itself is an inflammatory stimulus, there are many examples in which hypoxia signaling induces anti-inflammatory and tissue-protective responses, including that involving extracellular adenosine.

The Hypoxia-Adenosine Link

Similar to inflammation, an essential link exists between hypoxia, HIFs, and adenosine. As indicated earlier, it is well established that conditions of low oxygen or inflammation favor the release of extracellular ATP/ADP (48, 66). Indeed, cell membranes are damaged in these conditions, cell death is frequent, and nonlytic routes of ATP/ADP escape are exploited. Accordingly, extracellular adenosine levels increase at the cell surface. Considering that extracellular adenosine protects tissues in low oxygen, it was considered that adenosine metabolism and signaling genes were regulated by HIFs. Studies by the laboratory of Colgan (128) identified a HRE in the promoter of the CD73 gene and demonstrated CD73 to be a transcriptional target of HIF-1α. In 2004, with the successful generation of CD73-deficient mice, Thompson and colleagues (135) showed the extent to which CD73 gene regulation by HIF-1α was critical in low oxygen conditions. CD73-deficient mice suffer substantial vascular leakage and inflammatory cell accumulation in multiple organs with short-term exposure to low oxygen (49, 135). Extensions of these studies including CD39-deficient mice showed that extracellular adenosine production during hypoxia is immediate to the anti-inflammatory actions of neutrophils (49). CD39 is induced in hypoxia by transcription factor Sp1 (47). Sp1 is strongly implicated in hypoxic gene transcription and has been shown to play a protective role in regulation of CD39 during cardiac and hepatic ischemia (47, 59, 100).

CD73 is not the only adenosine pathway gene targeted by HIFs. Adenosine receptors, in particular A2AAR and A2BAR, are under the transcriptional control of HIFs (3, 76). Experimental studies provide evidence that A2AAR is a target gene of HIF-2α (3), whereas A2BAR is gene target of HIF-1α (76). Inasmuch as A2BAR protects the barrier in inflammation, pharmacological and genetic studies have demonstrated A2BAR is charged with protecting the barrier in low oxygen conditions. Indeed, A2BAR-deficient mice experience significant barrier leakage and increased tissue accumulation of immune cells when exposed to hypoxia (33). Genetic deletion studies of each adenosine receptor in mice indicate A2BAR as having a selective role in barrier protection in hypoxia (33). A2BAR expression is increased in murine studies of ischemia-reperfusion injury (60). As well, mice with gene deletion of CD73 and A2BAR experience more profound tissue injury with ischemia (41, 60). Ischemia results in a robust increase in HIFs, which is important to protecting tissues (40). Tissues experience significant destruction in the absence of HIF-1α with ischemia, whereas PHD2 absence substantially increases HIF-1α and attenuates tissue damage. Notably, HIF-1α-dependent ischemic protection is abolished in A2BAR-deficient mice (40). Studies with HIF activator, dimethyloxalylglycine (DMOG), provide another example in which the anti-inflammatory effects of HIF are routed through extracellular adenosine/adenosine signaling. Treatment of CD73 and A2BAR-deficient mice with DMOG provides no protection against ischemic injury (40).

HIFs have also been implicated in attenuating the termination of adenosine signaling. As such, HIFs directly repress the expression of ENT1 and ENT2 (43, 99) as well as reduce the intracellular conversion of adenosine to AMP via the adenosine kinase (98). Studies involving radiolabeled adenosine showed extracellular adenosine uptake slows in hypoxia and that ENT1 and ENT2 expression significantly decreases (43, 99). ENT1 and ENT2 are the main adenosine transporters that orchestrate the movement of adenosine from the cell surface to the intracellular space in vascular endothelial and epithelial cells. Accordingly, these studies demonstrated that ENT1 and ENT2 are direct transcriptional targets of HIF-1α, HIF-1α decreases ENT1 and ENT2 expression in hypoxia, and ENT1 and ENT2 repression attenuates inflammatory responses (43, 99). Interestingly, epithelial-targeted HIF-1α-deficient mice have increased expression of ENTs (43, 99). An elegant study by Schrader and colleagues (30) has shown that hypoxia-induced inhibition of adenosine kinase (AK) increases the release of extracellular adenosine. Studies with promoter constructs and a combination of HIF loss- and gain-of-function studies indicated AK is transcriptionally repressed by HIF-1α and inhibiting AK in vivo attenuates hypoxia-induced vascular leakage (98), thereby showing another adenosine pathway member targeted by HIF. Consistent with the role of HIFs in inducing extracellular adenosine and adenosine signaling, netrin-1 is increased by HIF-1α (119). As indicated earlier, netrin-1 can enhance A2BAR signaling to attenuate inflammatory responses (1, 96, 119). These studies show that by engaging A2BAR on neutrophils, HIF-1α-induced netrin-1 inhibits neutrophil transmigration, therefore dampening hypoxia-induced inflammation. Taken together, the above studies demonstrate that in hypoxia, HIFs coordinate the transcriptional control of multiple members of the adenosine pathway. HIFs increase extracellular adenosine-metabolizing enzymes and adenosine receptors, whereas expression of adenosine transport and intracellular metabolism genes are decreased. As such, this provides substantial support for adenosine signaling to protect tissues during conditions of low oxygen (Fig. 1). Additionally, studies involving erythrocyte metabolism and high-altitude adaptation have indicated an important role of the adenosine/purinergic pathway in driving adaptations to hypoxia. Exposure to hypoxia enhances erythrocyte glycolysis and higher ATP availability results in the accumulation of AMP and adenosine (24). Excellent studies by the laboratory of Xia (88) show soluble CD73 activity increases at high altitude and is associated with elevated erythrocyte 2,3-bisphosphoglycerate (2,3-BPG), a negative allosteric regulator of hemoglobin-O₂ binding affinity, the expression of which increases through erythrocyte A2BAR activity involving AMP-activated protein kinase. Here, A2BAR activation was beneficial in inducing 2,3-BPG and triggering O₂ release to prevent multiple tissue hypoxia, inflammation, and pulmonary leakage (88).

Examples for Hypoxia-Adenosine Link During Inflammatory Diseases

In the above paragraphs, we have discussed the associations of inflammation with adenosine, hypoxia with inflammation, and adenosine with hypoxia. Here, we focus on studies that bring together and emphasize the link between hypoxia and adenosine during inflammatory conditions and disease, specifically myocardial injury, inflammatory bowel disease, and acute lung injury. Together, these studies provide significant support that hypoxia-elicited, HIF-driven adenosine metabolism and signaling is protective to tissues, particularly by dampening inflammation and protecting tissue barriers.

Myocardial injury.

Myocardial injury is a significant health condition characterized by obstruction of blood supply to the heart. Damage occurs initially from the loss of blood supply and again with the sudden movement of blood back into cardiac tissues and reoxygenation. In many instances, profound inflammation occurs. Substantial evidence indicates that extracellular adenosine/adenosine signaling provides considerable cardioprotection during myocardial ischemia-reperfusion injury. For example, a robust induction of both CD39 and CD73 occurs in ischemic preconditioning. Deletion of CD39 or CD73 in mice results in larger infract sizes after closing off the coronary artery, whereas treatment of CD73-deficient mice with 5′-nucleotidase or CD39 mice with AMP or apyrase attenuates infarct sizes and provides additional cardioprotection benefit to wild-type mice during ischemic preconditioning (41, 75). The protection of CD73 to the heart has long been appreciated, as earlier studies in dogs revealed that inhibiting CD73 abolishes cardioprotection by ischemic preconditioning (73). Interestingly, the induction of CD39 expression by transcription factor Sp1 was shown to be central to providing cardioprotection in experimental studies of hypoxia and ischemia (47). Sp1 is strongly implicated in hypoxic gene transcription and shares similar target genes with HIFs (106).

Experimental studies in various models indicate that A1AR and A3AR are involved in providing protection before ischemia, A2AAR has promising protective actions during reperfusion, and more recent studies indicate that A2BAR is important as well in modulating myocardial reperfusion injury (95). Studies by Linden (141) indicated that A2AAR activity on T and B cells are involved in dampening ischemia-reperfusion injury in the heart. Other studies observed in a head-to-head comparison in mice deficient for each adenosine receptor that A2BAR-deficient mice are particularly susceptible to myocardial ischemia (41). Extension of these studies showed a robust increase in HIF-1α during ischemic preconditioning and that in vivo gene silencing of cardiac HIF-1α abolished cardioprotection, whereas pretreatment with HIF activator, DMOG, provided significant cardioprotection. Moreover, gene silencing of PHD2 resulted in significant HIF-1α activation and attenuated myocardial infarct sizes (40). Importantly, these studies demonstrated that HIF-dependent cardioprotection is through A2BAR. HIF-dependent cardioprotection was abolished in A2BAR-deficient mice (40). A2BAR providing cardioprotection is further supported by studies showing A2BAR agonist treatment significantly reduces infarct sizes after ischemia in wild-type mice but not A2BAR-deficient mice (41). More recently, the circadian rhythm protein period 2 (Per2) was identified as a target of HIF-dependent, A2BAR-elicited cardioprotection (36). A2BAR led to the stabilization of Per2 during myocardial ischemia, whereas Per2-deficient mice had larger infarct sizes. Ischemic hearts of Per2-deficient mice showed a limited ability to use carbohydrates for oxygen-efficient glycolysis, which was due to the stabilization of HIF-1α. The importance of the hypoxia-adenosine link in myocardial protection can be further supported by studies showing that hypoxia-induced inhibition of adenosine kinase potentiates cardiac adenosine release (30), and that ENT1-deficient mice are particularly more resilient to cardiac ischemia as opposed to wild-type mice (117). Further support includes studies of myocardial protection by remote ischemic preconditioning (RIPC) (112). RIPC is considered to protect the heart from subsequent ischemic myocardial injury. RIPC involves brief cycles of ischemia and reperfusion to an arm or a leg by inflating/deflating a cuff, which causes the release of protective molecules, such as adenosine. Myocardial protection by RIPC is mediated by adenosine receptor activity (112). Interestingly, the cardioprotection by RIPC is abrogated in HIF-1α-deficient mice and HIF-driven gene expression reduced (13).

Most of the above studies implicate adenosine and HIF in preventing myocardial injury during ischemic preconditioning. Certainly, the administration of adenosine or adenosine receptor antagonists, such as A2BAR agonist BAY 60-6583, or compounds that inhibit oxygen sensors, such as PHDs, thereby preventing HIF-α hydroxylation and leading to HIF stabilization, are attractive therapeutic approaches to minimizing myocardial injury. Both approaches would appear beneficial to administering to patients before undergoing major cardiac or thoracic surgery. PHD inhibitors are currently in active clinical trials for renal disease (Table 1) and appear to be safe for patients (8, 44, 51). The efficacy and safety of BAY 60-6583 in humans is unknown and remains mostly used in the context of experimental disease models (33, 35, 40, 57, 60, 121). Intravenous infusion of BAY 60-6583 is reported to have negligible effects on hemodynamics (82). Two multicenter clinical trials, Acute Myocardial Infarction Study of Adenosine (AMISTAD) I and II (91, 120), have examined the therapeutic potential of adenosine in the context of cardioprotection during reperfusion injury. Intravenous infusion of high-dose adenosine was shown to reduce myocardial infarct sizes. Patients administered adenosine with reperfusion therapy within the first few hours had significantly improved outcomes compared with placebo treatment (74). Importantly, these studies indicate that if adenosine reperfusion is accomplished within a small time period of onset of injury, the benefit of cardioprotection of adenosine is much greater. If reperfusion is delayed, it becomes rather difficult for adenosine to salvage the tissue. Therefore, the use of BAY 60-6583 or PHD inhibitors in situations such as a heart attack may have the greatest benefit when given in a defined time period during or after myocardial injury. Further experimental studies are needed to define such a therapeutic window. However, there is much hope for these compounds for the treatment of myocardial injury. Notably, adenosine has a considerably shorter half-life (1–2 s). Therefore, the use of BAY 60-6583 or PHD inhibitors, which are far more stable, would be considered to improve efficacy and reduce side effects due to adenosine’s nonselectivity [e.g., bradycardia, hypotension, rapid receptor desensitization (41)].

Table 1.

Therapeutic targets of the adenosine and hypoxia pathways and health conditions and diseases of current clinical trials

Drug	Target	Phase	Condition	Indications	https://ClinicalTrials.gov Identifier (Status)
ATP		Phase 2	Major depressive disorder	Antidepressant effect	NCT03138681 (recruiting)
Acetogenins	ATP production inhibitor	Phase 2/Phase 3	Breast cancer	Anti-proliferative, apoptosis effect	NCT02286778 (active, not recruiting)
Ticagrelor	P2Y12 antagonist	Phase 4	Microvascular angina	Ameliorate coronary microvascular function	NCT02284048 (not yet recruiting)
		Phase 4	Migraine headache	Eliminate migraine headaches	NCT02518464 (recruiting)
		Phase 3	Acute ST-elevation myocardial infarction	Antiplatelet therapy	NCT01739556 (recruiting)
		Phase 3	Diabetes mellitus, Type 2	Effect on cardiovascular death, myocardial infarction, or stroke	NCT01991795 (active, not recruiting)
Cangrelor	P2Y12 antagonist	Phase 1	Partial obstruction of systemic to pulmonary artery shunt; complete obstruction of systemic to pulmonary artery shunt	Safety profile in neonatal participants at risk of thrombosis	NCT02765633 (recruiting)
Cangrelor; Ticagrelor	P2Y12 antagonist	Phase 4	Acute coronary syndrome; microvascular obstruction; ST-segment elevation myocardial infarction; thrombolysis in myocardial infarction; unstable angina	Intravenous antiplatelet therapy	NCT02733341 (recruiting)
Clopidogrel; Ticagrelor	P2Y12 antagonist	Phase 4	Coronary artery disease	Reduce later heart attacks	NCT02327624 (active, not recruiting)
		Phase 4	Patients who are scheduled to undergo a PCI (percutaneous coronary intervention) for CTO (chronic total occlusion)	Percutaneous coronary intervention and the recovery of vascular function at long-term	NCT02211066 (recruiting)
		Phase 3	Acute coronary syndrome	Prevention of fatal and nonfatal cardiovascular events, observation of coronary circulation effects due to chronic use	NCT02618733 (recruiting)
		Phase 2	Coronary artery disease; kidney dysfunction	Reducing ischemic events	NCT03039205 (not yet recruiting)
		Phase 2	Stroke; ischemic attack, transient	Combination with aspirin effect on reducing transient ischemic attack or minor stroke	NCT02506140 (recruiting)
Prasugrel; Ticagrelor	P2Y12 antagonist	Phase 4	Acute coronary syndrome	Pleiotropic effects to antiplatelet therapy in type 2 diabetic patients with non-ST elevation acute coronary syndrome	NCT02487732 (active, not recruiting)
MEDI9447 with or without MEDI4736	CD73 (MEDI9447); anti-PD-LI antibody (MEDI4736)	Phase 1	Solid tumors	Safety profile in adult subjects with select advanced solid tumors (combat tumor immunosuppression)	NCT02503774 (recruiting)
Adenosine		Phase 2	Cardiomyopathies	An adjunct to intermittent warm blood cardioplegia	NCT02681913 (recruiting)
		Phase 2	Takotsubo cardiomyopathy; Takotsubo syndrome	Induce rapid recovery of left ventricle function	NCT02867878 (recruiting)
		Phase 2/Phase 1	Sinus bradycardia; atrioventricular block	Safety profile in children with heart transplant, effectiveness in treating fast heart rate	NCT02462941 (active, but not recruiting)
Neladenoson Bialanate (BAY 1067197)	A1AR agonist	Phase 2	Heart failure	Optimal dose given in addition to standard therapy for heart failure	NCT02992288 (recruiting)
Regadenoson	A2AAR agonist	Phase 2	Retinal artery occlusion	Induce vasodilatation in adjacent retinal areas to increase blood flow and retinal oxygenation	NCT03090087 (recruiting)
		Phase 2	Sickle cell anemia	Effective treatment for pain crises and acute chest syndrome in sickle cell disease	NCT01788631 (recruiting)
		Phase 1	Lung transplant	Ischemia reperfusion injury lowered or prevented. Regadenoson is an A2AAR drug	NCT03072589 (not yet recruiting)
		Phase 1	Blood-brain barrier defect	Transiently disrupt the blood-brain barrier, improve drug penetration	NCT02389738 (recruiting)
		N/A	Sickle cell disease; sickle cell anemia	Prevent inflammation and injury caused by the sickle shaped cells	NCT01566890 (recruiting)
CF102	A3AR agonist	Phase 2	Nonalcoholic steatohepatitis	Safety and efficacy, improve liver pathology	NCT02927314 (not yet recruiting)
		Phase 2	Hepatocellular carcinoma	Efficacy and safety, tumor reduction, improved survival	NCT02128958 (recruiting)
Aminophylline	Nonselective adenosine receptor (AR) antagonist	Phase 1	Premature obstetric labor	Combination with progesterone reduces the risk of preterm labor	NCT03152942 (not yet recruiting)
		Phase 1	Acute kidney injury	Mitigates this vasoconstriction, reduce ischemic kidney injury	NCT02983422 (not yet recruiting)
Vivarin (caffeine)	Nonselective AR antagonist	Phase 2/Phase 1	Narcolepsy	Effects of caffeine consumption on daytime sleepiness and reaction time	NCT02832336 (recruiting)
Caffeine	Nonselective AR antagonist	Phase 2/Phase 1	Pain	Assess caffeine on acupuncture analgesia	NCT02577770 (recruiting)
		N/A	Schizophrenia	Improve cognitive function	NCT02832401 (not yet recruiting)
PBF-680	A1AR antagonist	Phase 2	Asthma	To attenuate “late asthmatic responses” as a primary efficacy outcome, assess airway inflammation-related outcomes	NCT02635945 (recruiting)
		Phase 3	Acute coronary syndrome	Prevention of fatal and nonfatal cardiovascular events, observation of coronary circulation effects due to chronic use	NCT02618733 (recruiting)
Istradefylline	A2AAR antagonist	Phase 3	Idiopathic Parkinson’s disease	Safety and tolerability, nondopaminergic therapy	NCT02610231 (active, not recruiting)
Pentoxifylline	A2AAR antagonist	Phase 3	Acute pancreatitis; gallstone pancreatitis; alcoholic pancreatitis; Post-ERCP/postprocedural pancreatitis; trauma acute pancreatitis; hypertriglyceridemia acute pancreatitis; idiopathic (unknown) acute pancreatitis; medication-induced acute pancreatitis; cancer acute pancreatitis; miscellaneous (i.e., acute or chronic pancreatitis)	Inhibition of the tumor necrosis factor-alpha pathway reduces inflammation	NCT02487225 (enrolling by invitation)
PBF-509 with or without PDR001	A2AAR antagonist (PBF-509); anti-PD-1 antibody (PDR001)	Phase 2/Phase 1	Non-small cell lung cancer (NSCLC)	Safety, tolerability, feasibility, and preliminary efficacy (combat tumor immunosuppression)	NCT02403193 (recruiting)
CPI-444 with or without atezolizumab	A2AAR antagonist (CPI-444); anti-PD-L1 antibody (atezolizumab)	Phase 1	NSCLC; malignant melanoma; renal cell cancer; triple negative breast cancer; colorectal cancer; bladder cancer	Safety, tolerability, and anti-tumor activity (combat tumor immunosuppression)	NCT02655822 (recruiting)
EZN-2279; Adagen	Adenosine deaminase (ADA)	Phase 3	ADA-SCID; adenosine deaminase deficiency; severe combined immunodeficiency	Safety, efficacy, and pharmacokinetics [enzyme (ADA) replacement therapy]	NCT01420627 (recruiting)
ADA gene transduced CD34+ cells	ADA	Phase 2	Immunologic deficiency syndromes	Safety and the clinical efficacy of gene therapy (ADA), evaluate the immunological reconstitution and purine metabolism after gene therapy	NCT00598481 (active, not recruiting)
EF1αS-ADA	ADA	Phase 2/Phase 1	Adenosine deaminase deficiency; severe combined immunodeficiencies (SCID)	Safety, feasibility, effectiveness of gene therapy for ADA-deficient SCID using lentiviral vector	NCT01380990 (recruiting)
EFS-ADA	ADA	Phase 2/Phase 1	ADA-SCID	safety and effectiveness of using a lentiviral vector (based on HIV-1) for ADA-deficient SCID	NCT01852071 (active, not recruiting)
Pentostatin	ADA inhibitor	Phase 3	Graft-vs.-host disease	Efficacy of ruxolitinib against best available therapy in participants with steroid-refractory chronic graft-vs.-host disease	NCT03112603 (recruiting)
		Phase 2	Renal cell carcinoma; Graft-vs.-Host disease; Engraftment Syndrome	Determine whether new, low-intensity transplant approach can yield objective partial or complete remission	NCT00923845 (active, not recruiting)
		Phase 2	Leukemia; lymphoma	Stop the growth of cancer cells, combination therapy	NCT00602836 (active, not recruiting)
		Phase 2	Acute lymphoblastic leukemia; acute myeloid leukemia; chronic lymphocytic leukemia; chronic myelogenous leukemia, BCR-ABL1 positive; graft vs. host disease; Hodgkin lymphoma; myelodysplastic/myeloproliferative neoplasm; non-Hodgkin’s lymphoma; plasma cell myeloma; Waldenstrom macroglobulinemia	Preventing graft rejection in patients who have undergone donor stem cell transplant	NCT00096161 (active, not recruiting)
		Phase 2	Hairy cell leukemia	Determine whether pentostatin or bendamustine is a more effective when combined with rituximab	NCT01059786 (recruiting)
		Phase 2	Graft-vs.-host disease	Effectiveness in treating chronic graft-vs.-host disease in patients refractory to treatment with steroids	NCT00074035 (active, not recruiting)
		Phase 2/Phase 1	Sickle cell disease	Novel immunosuppressive regimen without myeloablation to further decrease the transplant-related morbidity/mortality	NCT03077542 (recruiting)
		Phase 2/Phase 1	Sickle cell disease; thalassemia; stem cell transplantation; graft vs. host disease	Reduce the transplant failure rate	NCT02105766 (recruiting)
		Phase 2/Phase 1	Mesothelioma; adenocarcinoma of lung; pancreatic neoplasms	Safety, tolerability, and feasibility of recombinant anti-mesothelin immunotoxin in combination with immune-depleting regimen	NCT01362790 (active, not recruiting)
		Phase 1	Leukemia, B cell; lymphoma, Hodgkin’s lymphoma; non-Hodgkin’s lymphoma, B cell	Preventing graft rejection	NCT01087294 (recruiting)
		Phase 1	Chronic lymphocytic leukemia; B cell non-Hodgkin's lymphoma	Synergy between bendamustine and purine analogs in killing cancer cells	NCT01352312 (active, not recruiting)
Vadadustat (AKB-6548)	Hypoxia-inducible factor-prolyl hydroxylase (PHD) inhibitor	Phase 3	Anemia; non-dialysis-dependent chronic kidney disease	Correction of anemia and maintenance of hemoglobin	NCT02648347 (recruiting)
		Phase 3	Anemia; dialysis-dependent chronic kidney disease	↓	NCT02865850 (recruiting)
		Phase 3	Anemia; non-dialysis-dependent chronic kidney disease	↓	NCT02680574 (recruiting)
		Phase 3	Anemia in subjects With DD-CKD	↓	NCT02892149 (recruiting)
		Phase 2	Anemia; dialysis-dependent chronic kidney disease	↓	NCT03054350 (active, not recruiting)
		Phase 2	Anemia; non-dialysis dependent chronic kidney disease	↓	NCT03054337 (active, not recruiting)
		Phase 2	Anemia; dialysis-dependent chronic kidney disease	↓	NCT03140722 (recruiting)
Daprodustat (GSK1278863)	PHD inhibitor	Phase 3	Anemia	Maintain hemoglobin	NCT03029208 (recruiting)
		Phase 3	Anemia	↓	NCT02969655 (recruiting)
		Phase 3	Anemia	↓	NCT02876835 (recruiting)
		Phase 3	Anemia	↓	NCT02879305 (recruiting)
		Phase 2	Anemia	↓	NCT03029247 (not yet recruiting)
Roxadustat (FG-4592); also known as AP1517	PHD inhibitor	Phase 3	Anemia in chronic kidney disease in nondialysis patients	To treat anemia	NCT02021318 (recruiting)
		Phase 3	Anemia	↓	NCT02174627 (active, not recruiting)
		Phase 3	Anemia; end-stage renal disease (ESRD)	↓	NCT02278341 (active, not recruiting)
		Phase 3	Anemia	↓	NCT02174731 (recruiting)
Roxadustat (FG-4592); also known as AP1517	PHD inhibitor	Phase 3	Anemia in chronic kidney disease in nondialysis patients	To treat anemia	NCT01887600 (active, not recruiting)
		Phase 3	CKD anemia in stable dialysis patients	↓	NCT02273726 (recruiting)
		Phase 3	Anemia in incident dialysis patients	↓	NCT02052310 (recruiting)
		Phase 1	Normal renal function; impaired renal function	Evaluate pharmacokinetics in subjects with different degrees of renal function	NCT02965040 (recruiting)
ASP1517	PHD inhibitor	Phase 3	Chronic kidney disease	Oral dosing; treatment of anemia	NCT02988973 (recruiting)
		Phase 3	Peritoneal dialysis chronic kidney disease patients with anemia	↓	NCT02780726 (active, not recruiting)
		Phase 3	Chronic kidney disease	↓	NCT02964936 (recruiting)
		Phase 3	Hemodialysis chronic kidney disease patients With anemia	↓	NCT02952092 (recruiting)
Remote Ischemic Preconditioning		N/A	Myocardial infarction	Effective endogenous cardiac protection, assess major adverse cardiovascular events such as cardiovascular death, spontaneous myocardial infarction, unplanned revascularization and stroke	NCT02843464 (not yet recruiting)
		N/A	Myocardial infarction, acute	Effective endogenous cardiac protection	NCT03018873 (not yet recruiting)

↓, Same as previous indication.

Inflammatory bowel disease.

Inflammatory bowel disease (IBD), which includes Crohn’s disease and ulcerative colitis, represents a group of intestinal disorders caused by dysregulated immune responses and aberrant mucosal barrier protection. IBD is characterized by excessive inflammation of the bowel and profound hypoxia (17, 71). A substantial number of studies show HIFs are stabilized in experimental models of colitis (23, 71, 99, 116) and have almost uniformly beneficial influence on disease outcomes (16). For example, mice with deletions of HIF-1α in intestinal epithelial cells are more susceptible to intestinal inflammation and have more severe disease during experimental colitis (16, 44, 71, 132), whereas deletion of HIF-2α appears to be protective (140). HIF-1α deletion was associated with increased weight loss, increased intestinal inflammation, and more profound colonic shortening (71). Similarly, the deletion of VHL (71) or PHD1 (129) in mice appears to be protective during intestinal inflammation, again implicating HIFs in gut protection during IBD (16, 44, 132). Notably, CD73 expression was profoundly induced in VHL-deficient mice (71). Several studies demonstrate that pharmacological compounds that stabilize HIFs provided potent protection during intestinal inflammation (23, 44, 52, 116). In general, CD39- (54) and CD73 (89)-deficient mice develop increased disease severity during experimental colitis. For example, CD73-deficient mice experienced significant weight loss and colonic shortening with trinitrobenzene sulfonate (TNBS) colitis, and similarly, treatment of wild-type mice with CD73 inhibitor α,β-methylene ADP results in a similar increase in colitis severity (89). The onset of significant gut hypoxia during IBD was shown by the laboratory of Colgan (16, 71, 132) using murine IBD models and 2-nitroimidazole compounds. In addition, specimens from patients with IBD show prominent HIF-1α and HIF-2α activation (55). Similar to CD73 (128), A2BAR is directly induced by HIF-1α (76). Several studies implicate A2BAR signaling in IBD (1, 2, 53). A2BAR-deficient mice have an increased susceptibility to dextran sulfate sodium (DSS)-induced colitis; mice experience increased weight loss, colonic shortening, and disease activity indexes. Similarly, severity of colitis is increased in wild-type mice treated with the A2BAR antagonist PSB1115 (53). A2BAR-deficient mice as well as treatment of intestinal epithelial cells with PSB1115 show deficiencies in the anti-inflammatory cytokine IL-10, implicating A2BAR signaling as having anti-inflammatory activity in IBD. Both CD73 and A2BAR-deficient mice suffer extensive breakdown of tissue barriers with hypoxia (10, 33, 135). Studies have shown intestinal epithelial A2BAR provides significant protection during experimental colitis by enhancing barrier protection responses.

Indeed, mice with global or tissue-specific deletion of A2BAR experience increased severity of colitis, associated with a more acute onset of disease and epithelial barrier breakdown (1, 2, 53). In vitro and in vivo studies implicated A2BAR-driven phosphorylation of vasodilator-stimulated phosphoprotein (VASP) as a specific response to regain the barrier (2). These studies showed as well epithelial A2BAR signaling to selectively attenuate gut inflammation as opposed to vascular endothelial A2BAR (2). As well, other studies have demonstrated netrin-1-deficient mice to suffer significant disease severity during DSS colitis (1). Previous studies show HIF-1α-dependent induction of netrin-1 to attenuate hypoxia-elicited inflammation of mucosal surfaces by netrin-1 engaging A2BAR on neutrophils (119). Netrin-1-deficient mice show increased weight loss and colonic shortening during experimental colitis, whereas mice treated with exogenous netrin-1 show far less disease severity. These studies further showed that netrin-1-dependent attenuation of colitis is through suppression of neutrophil trafficking by A2BAR (119). Altogether, these studies provide substantial evidence that HIF-1-elicited adenosine metabolism and signaling is an essential pathway that provides considerable anti-inflammatory responses and barrier production during IBD.

Several studies have indicated and more are ongoing that provide the basis for therapeutic use of PHD inhibitors in IBD (44, 116). A concern of PHD inhibitors is that they are nonspecific for PHD isoforms, and many tissues express PHDs. As well, PHD inhibitors have the potential to inhibit other hydroxylases. The development of tissue-specific delivery of PHD inhibitors has the potential to improve upon these concerns and is an area of rising interest. Experimental studies have shown significant benefit of PHD inhibitor, AKB-4924, in murine models of colitis (72). Specifically, these studies demonstrated PHD inhibition in gut epithelial cells is immediate to providing therapeutic benefit in colitis. Oral administration of AKB-4924 therefore may be ideal as opposed to systemic administration of PHD inhibitors. Such an approach may limit possible adverse side effects. The oral PHD inhibitor roxadustat (FG-4592) is in clinical trials for renal disease (Table 1). Experimental studies also support the therapeutic benefit of the A2BAR agonist BAY 60-6583 in IBD (2). It would be worthwhile to assess the possible use of oral administration of BAY 60-6583 as a therapeutic strategy for IBD.

Acute lung injury.

Acute lung injury (ALI) is a lung disease characterized by pulmonary edema and severe hypoxia (138). ALI is caused by injuries or acute infections to the lungs. The past decade has provided considerable evidence of a link between hypoxia and adenosine signaling in dampening pulmonary edema and promoting resolution of lung inflammation (34, 35, 37–39, 121, 138). Patients with ALI often require mechanical ventilation due to difficulties in breathing, whereas ALI often occurs with mechanical ventilation. Studies using cyclic mechanical stretch, a model that mimics cyclic opening and closing of the airway during mechanical ventilation, identified that stretched pulmonary epithelial cells release extracellular adenosine (34, 138). Indeed, mechanical ventilation in mice increases pulmonary extracellular adenosine. Here, studies showed CD39- and CD73-deficient mice suffer severe ALI with mechanical ventilation. Similarly, drug inhibition of CD73 in mice induces significant pulmonary edema and inflammation, whereas treatment of CD39- or CD73-deficient mice with apyrase or 5′-nucleotidase significantly attenuates ventilator-induced ALI, implicating the role of extracellular adenosine in lung protection during ALI (34). Consistent with the continuing theme of A2BAR signaling to protect the barrier, studies show A2BAR is specifically associated with attenuating ventilator-induced ALI (35). Both A2BAR-deficient and wild-type mice treated with an A2BAR antagonist suffer pulmonary edema, inflammation, and attenuated gas exchange, whereas an A2BAR agonist attenuates ALI. Additional studies showed pulmonary inflammation involves A2BAR signaling from bone marrow-derived cells, whereas A2BAR on the lung tissue is responsible for attenuating pulmonary edema. A2BAR is also implicated in the protection of ALI induced by endotoxins (121). Other studies implicate ENT-dependent adenosine uptake enhances A2BAR signaling on pulmonary epithelial cells to dampen ALI (37). These studies revealed ENT2 to be selectively involved in attenuating pulmonary edema and improving gas exchange during ALI. Notably, increased extracellular adenosine was found in bronchoalveolar fluid, and ENT-dependent lung protection was abolished in mice with epithelial A2BAR deletion.

Recent studies identified HIF-1α as a transcriptional inducer of A2BAR during ventilation-induced ALI (38). Stretch-induced A2BAR was abolished with silencing of HIF-1α in pulmonary epithelial cells. Similarly, drug inhibition of HIF-1α inhibited the increase of A2BAR expression during cyclic stretch. Moreover, studies of ventilator-induced ALI revealed induction of A2BAR expression was abolished in conditional HIF-1α-deficient mice. An elegant study from the laboratory of Sitkovsky (134) considered that oxygenation may weaken hypoxia-elicited A2AAR-mediated anti-inflammatory responses, thereby further exacerbating ALI. The study is based on the observation that hypoxia induces A2AAR signaling that attenuates lung inflammation and tissue damage. In a bacterial infection ALI model, wild-type mice exposed to 100% oxygen, mimicking therapeutic oxygenation, had reduced survival rates compared with mice exposed to 21% oxygen. Indeed, five times more mice died after exposure to 100% oxygen. Intratracheal injection of an A2AAR agonist mimics the protective effects of hypoxia upon oxygenation with ALI, thereby decreasing lung injury and improving survival rates. Thus high oxygen percentages, as provided to patients suffering from ALI, may weaken hypoxia-driven, adenosine signaling-mediated protection of the lungs. Another study implicates HIF-2α in the control of the A2AAR in human pulmonary endothelial cells during hypoxia (3). Taken together, these studies emphasize that hypoxia is protective in ALI and involves the stabilization of HIFs and HIF-directed induction of adenosine metabolism and signaling.

Similar to IBD, patients with ALI are likely to benefit from therapeutic strategies that increase HIF stabilization and adenosine signaling, especially therapeutic approaches that are specifically directed to the lung. Recent studies have shown that inhaled BAY 60-6583 treatment during ALI in mice attenuates pulmonary edema, reduces inflammation, and protects against tissue damage of the lungs (63). Here, these studies found A2BAR of alveolar epithelial cells are central to lung protection during ALI. Inhaled drug delivery has a significant benefit to systemic delivery in that systemic side effects would be reduced and lower concentration of the drug could be used. PHD inhibitors have shown significant increases in survival and decreased lung inflammation during mechanical ventilation-induced ALI (32). These studies also implicate HIF-1α stabilization specifically in alveolar epithelial cells to be central to dampening lung inflammation during ALI. The delivery of PHD inhibitors by an inhaled approach may also provide significant benefit during ALI. However, future studies are needed to determine the benefit of such an approach.

Conclusions

Great progress has been made over the past decade to bring the adenosine pathway to the clinic. Indeed, many adenosine and several adenosine receptor agonists and antagonists are currently in clinical trials or are in use in preclinical studies in many conditions and diseases, such as Parkinson’s disease, sickle cell anemia, and cancer (Fig. 2 and Table 1). Recently completed clinical trials for adenosine metabolism and signaling are summarized in Fig. 3. In the present review, we discussed that adenosine metabolism and signaling is essential to dampening inflammation and protecting tissue barriers, and immediately involved with this are HIFs. Approaches to therapeutically target this response in myocardial injury, IBD, or ALI possibly include drugs that decrease proinflammatory responses induced through ATP/ADP receptors, such as P2Y12. Ticagrelor and Clopodogrel, both P2Y12 antagonists, are in clinical trials for several health conditions. Drugs that increase extracellular adenosine, such as adenosine analogs, dipyridamole, an ENT inhibitor, or pentostatin, an ADA inhibitor, may also have benefit. Pentostatin is currently in clinical trials for graft vs. host disease and different cancers; it is FDA approved for the treatment of hairy cell leukemia. Agents that increases A2AAR and A2BAR signaling serve as possible treatment targets as well, such as Regadenoson, an A2AAR agonist, and BAY 60-6583. Regadenoson is in clinical trials for a number of heart conditions as well as sickle cell disease. BAY 60-6583 use remains in experimental settings (33, 35, 40, 57, 60, 121). Adenosine receptor antagonists and CD73/CD39 inhibitors are being studied in clinical trials as potent anticancer drugs, specifically as agents that boost immune responses against tumors (4). On the contrary, increasing adenosine levels show benefit in certain leukemia and lymphomas and are related to adenosine inducing apoptosis in the cancer cells (115). Similarly, CF-102 (A3R agonist) has shown promise to patients with hepatocellular cancer in recent clinical trials and is suspected of inducing cancer cell apoptosis (6, 125). Adenosine metabolism and signaling can have very different roles in different tumors and cells of the tumor microenvironment (10, 11).

Fig. 2.

Therapeutic targets of the adenosine and hypoxia pathways and health conditions and diseases of current clinical trials. Left: multiple drugs exist for targeting adenosine metabolism and signaling and HIF activation. Many are in clinical trials (right panel and Table 1) or in use in preclinical studies. Dampening inflammation and protecting tissues barriers can be achieved by approaches that decrease proinflammatory responses (1), increase extracellular adenosine (2, 3, 5, 6), and increase adenosine signaling (2–6). Examples of drugs and drug targets include 1) Ticagrelor and Clopodogrel, P2Y12 (ATP/ADP receptor) antagonists; 2) DMOG and FG-4497, HIF stabilizers (PHD inhibitors); 3) adenosine and adenosine analogs [e.g., 5′-(N-ethylcarboxamido)], nonselective adenosine receptor agonists; 4) Regadenoson and BAY 60-6583, A2AAR and A2BAR agonists, respectively; 5) dipyridamole, ENT inhibitor; and 6) Pentostatin, ADA inhibitor. Adenosine receptor antagonists (e.g., A2AAR antagonists, PBF-509 and CPI-444) or CD73/CD39 inhibitors (e.g., MEDI9447) are currently in clinical trials in a number of solid tumors and are proposed as anticancer therapies that boost immune responses against tumor cells. On the contrary, Pentostatin, which increases adenosine levels, is currently in clinical trials for different solid tumors, leukemia, and lymphoma. In a similar manner, CF-102 an A3AR agonist has been shown to have therapeutic potential in hepatocellular cancer.

Fig. 3.

Summary of recently conducted clinical trials targeting adenosine metabolism and signaling.

Understanding the molecular basis of oxygen sensing has proven to have great implications for medicine. Recently, PHD inhibitors have made their way into active clinical trials for kidney disease; these agents are appearing to have great promise (44) (Table 1). Alternatively, given that HIFs increase CD73, A2AAR, and A2BAR expression, while decreasing ENT and AK expression, thereby increasing adenosine metabolism and signaling, the use of HIF stabilizers, such as PHD inhibitors (e.g., FG-4497), may serve as a promising therapeutic approach to targeting the adenosine pathway in inflammatory diseases. Adenosine signaling and PHDs have significant involvement in other tissues, which could produce limiting adverse side effects. Further advancement of cell- or tissue-specific delivery approaches of these agents may make all the difference in the future success of these therapies. As well, introducing more clinical trials that take into account the various challenges that have hampered the success of prior adenosine pathway drugs are certainly needed (14). We are optimistic for the future of these therapeutic targets and their potential as treatments of inflammatory conditions and diseases.

GRANTS

This research was supported by an International Anesthesia Research Society Mentored Research Award to J. L. Bowser, National Institutes of Health (NIH) Grants P50-CA-098258 and DK-056338 to J. L. Bowser, and NIH Grants R01-DK-097075, R01-HL-098294, POI-HL-114457, R01-DK-082509, R01-HL-109233, R01-DK-109574, R01-HL-119837, and R01-HL-133900 to H. K. Eltzschig.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.L.B. prepared figures; J.L.B. drafted manuscript; J.L.B., J.W.L., X.Y., and H.K.E. edited and revised manuscript; J.L.B., J.W.L., X.Y., and H.K.E. approved final version of manuscript.