Retracted: Equilibrative nucleoside transporter (ENT)-1-dependent elevation of extracellular adenosine protects the liver during ischemia and reperfusion

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Ischemia and reperfusion-elicited tissue injury contributes to morbidity and mortality of hepatic surgery and during liver transplantation. Previous studies implicated extracellular adenosine signaling in liver protection. Based on the notion that extracellular adenosine signaling is terminated by uptake from the extracellular towards the intracellular compartment by way of equilibrative nucleoside transporters (ENTs), we hypothesized a functional role of ENTs in liver protection from ischemia. During orthotopic liver transplantation in humans, we observed higher expressional levels of ENT1 than ENT2, in conjunction with repression of ENT1 and ENT2 transcript and protein levels following warm ischemia and reperfusion. Treatment with the pharmacologic ENT inhibitor dipyridamole revealed elevations of hepatic adenosine levels and robust liver protection in a murine model of liver ischemia and reperfusion. Studies in gene-targeted mice for Ent1 or Ent2 demonstrated selective protection from liver injury in Ent1−/− mice. Treatment with selective adenosine receptor antagonists indicated a contribution of Adora2b receptor signaling in ENT-dependent liver protection. Conclusion: These findings implicate ENT1 in liver protection from ischemia and reperfusion injury and suggest ENT inhibitors may be of benefit in the prevention or treatment of ischemic liver injury. (Hepatology 2013;58:1766–1778)

Abbreviations
ADP

adenosine diphosphate

AMP

adenosine monophosphate

AR

adenosine receptor

ATP

adenosine triphosphate

cAMP

cyclic adenosine monophosphate

CIT

cold ischemia time

ENT

equilibrative nucleoside transporter

Ischemia and reperfusion is a pathologic condition characterized by an initial restriction of blood supply to an organ, followed by the subsequent restoration of perfusion and concomitant reoxygenation.[1, 2] In its classic manifestation, occlusion of the arterial blood supply is caused by an embolus and results in a severe imbalance of metabolic supply and demand causing tissue hypoxia. In the second stage of the disease, blood flow is rapidly restored. Somewhat surprisingly, the restoration of blood flow along with reoxygenation is frequently associated with an exacerbation of tissue injury and a profound inflammatory response (so-called reperfusion injury).[3] While ischemia and reperfusion contribute significantly to a wide range of pathologies, its functional contribution during liver surgery is particularly severe. For example, ischemia and reperfusion is a frequent cause of acute liver failure during orthotopic liver transplantation. Similarly, ischemia and reperfusion injury can contribute to immunologic consequences during human liver transplantation, as it is implicated in early rejection of the transplanted liver graft or the recurrence of hepatitis C in patients undergoing liver transplantation for the treatment of chronic hepatitis. Moreover, treatment modalities that would prevent hepatic ischemia and reperfusion injury are very limited and studies that aim to identify novel therapeutic approaches for hepatic ischemia and reperfusion are an area of intense investigation.[4, 5]

Previous studies had shown that ischemia and reperfusion is associated with increased adenosine production from its precursor molecules—particularly the nucleotides adenosine triphosphate (ATP), adenosine diphosphate (ADP), and adenosine monophosphate (AMP).[1, 6] Furthermore, it has been shown that activation of cyclic adenosine monophosphate (cAMP)-dependent protein kinase A regulates local inflammation and prevents hepatocyte death.[7] Extracellular adenosine can signal through four distinct adenosine receptors (ARs), Adora1, Adora2a, Adora2b, or Adora3.[1] Studies of hepatic ischemia and reperfusion had shown a functional role for extracellular adenosine production,[8, 9] and signaling events through ARs—such as Adora2a[10] and Adora2b[11]—in liver protection from ischemia. Based on the notion that extracellular adenosine signaling is terminated by way of uptake of adenosine from the extracellular towards the intracellular compartment by way of equilibrative nucleoside transporters (ENTs),[12-15] we pursued the hypothesis that inhibition of extracellular adenosine uptake could represent a means of enhancing hepatic adenosine signaling and concomitant liver protection from ischemia. Indeed, by attenuating adenosine uptake, and concomitant increases of spontaneously formed extracellular adenosine levels, endogenously generated levels of extracellular adenosine could become sufficient to trigger immunosuppressive adenosine receptor signaling events within the inflamed liver tissue microenvironments in vivo.[16, 17] In other words, the molecular concept is that pathophysiologically induced elevations of extracellular adenosine can provide better liver protection, if they are elevated over a longer time period, which can be achieved by adenosine uptake inhibitors, or by genetically targeting individual adenosine transporters. Therefore, we combined studies of ENT transcript and protein levels in biopsy samples obtained from patients undergoing liver transplantation with pharmacologic and genetic studies in a previously described model of murine partial hepatic ischemia and reperfusion.[8, 9, 18] These studies demonstrated a selective role for ENT1 in elevating hepatic adenosine levels and conveying liver protection from ischemia and reperfusion injury.

Materials and Methods

Human Liver Tissue

Liver samples were obtained from patients undergoing orthotopic liver transplantation (Supporting Table 1). Liver biopsies (I) were taken at the conclusion of cold ischemia time (CIT) during back table preparation of the cadaveric liver allograft (Fig. 1A). A second biopsy (R) was taken immediately prior to closure of the abdomen following drain placement (Fig. 1A). Importantly, total reperfusion time is defined as the time from portal vein perfusion to abdominal closure at the conclusion of the procedure.

Figure 1.

ENT1 and ENT2 expression in biopsies of human liver transplants following hepatic ischemia and reperfusion. (A) First liver biopsy was taken during ischemia (I) at the conclusion of cold ischemia time (CIT) during back table preparation of the cadaveric liver allograft. A second biopsy was taken during reperfusion (R) after warm ischemia time (WIT) and reperfusion time (RT) immediately prior to closure of the abdomen following drain placement. (B) ENT1 and ENT2 transcript levels in biopsies of human liver transplants during cold ischemia (I) and after reperfusion (R); (n = 3-4 independent experiments). (C) ENT1 and ENT2 protein levels (one representative blot of two is shown).

Mice

All animal protocols were in accordance with the University of Colorado, Denver guidelines. Ent1 on the C57BL/6J strain were generated, validated, and characterized as described.[19] Ent2-deficient mice were obtained from Taconic Farms. Conditional hypoxia-inducible factor 1 alpha (HIF1α)loxP/loxP Albumin Cre+ mice were obtained by crossing HIF1αloxP/loxP with Albumin Cre+ mice (Jackson Laboratory). In all control experiments, age-, gender-, and weight-matched littermate controls were used.

Murine Model of Partial Liver Ischemia

In an effort to avoid mesenteric congestion, a murine model of partial liver ischemia was employed using a hanging-weight system as described.[18]

Transcriptional Analysis

Ent1 and Ent2 transcript levels were measured by reverse-transcription polymerase chain reaction (RT-PCR) (iCycler, Bio-Rad Laboratories) as described.[20]

Immunoblotting

In both human and mouse tissues Ent1 and Ent2 protein content was determined at different timepoints as described.[20]

Isolation of Hepatocytes

Liver preparation was performed as described in detail by Wei and colleagues.[21]

Enzyme-Linked Immunosorbent Assay (ELISA) (interferon-gamma [IFN-γ], interleukin 6 [IL-6], myeloperoxidase [MPO])

IFN-γ, IL-6 (R&D Systems), and neutrophil sequestration was quantified according to the manufacturer's instructions.

Adenosine Measurement

Livers were removed and immediately snap-frozen after 45 minutes of liver ischemia without reperfusion. Adenosine was measured as described.[22]

Liver Histology

Liver tissue was harvested following 2 or 24 hours of reperfusion. Sections (3 μm) were stained with hematoxylin and eosin (H&E). Examination and scoring (Suzuki Scoring 0-4) based on the presence and/or severity of sinusoidal congestion, cytoplasmic vacuolization, and necrosis of parenchymal cells was performed for six representative sections of each liver sample (n = 4-6 for each condition) in a blinded fashion.[9] Tissue injury was scored

Statistical Analysis

Liver injury score data are given as median and range. All other data are presented as mean ± SD from three to eight animals per condition. We performed statistical analysis using the Student t test. A value of P < 0.05 was considered statistically significant. For western blot analysis two to three repeats were performed. For all statistical analysis GraphPad Prism 5.0 software for Windows XP was used.

Study Approval

Collection and use of patient samples were approved by the COMIRB at UC Denver. All animal protocols were in accordance with the United States Guidelines Institutional Animal Care and Use Committee (IACUC) for use of living animals and were approved by the Institutional Animal Care and Use Committee of the University of Colorado guidelines for animal care.

Results

Human ENT Transcript and Protein Levels Are Repressed Following Orthotopic Liver Transplantation

Previous studies had indicated that termination of extracellular adenosine signaling is terminated by way of uptake of adenosine from the extracellular towards the intracellular compartment by way of ENTs.[12-15] Such studies also revealed that the transcriptional regulation of ENTs represents an important regulatory mechanism to alter adenosine signaling events. For example, transcriptional repression of ENTs during hypoxia results in enhanced extracellular adenosine accumulation and represents an endogenous antiinflammatory pathway to dampen hypoxia-induced inflammation.[12, 15] Along the lines of these studies, we pursued the hypothesis that ENTs could be important regulators of hepatic adenosine signaling during liver ischemia, thereby contributing to adenosine-dependent liver protection from ischemia. Therefore, we examined the expression of ENTs in human liver biopsy samples. We obtained biopsy samples during orthotopic liver transplantation, with the first biopsy taken following organ procurement and cold ischemia (baseline) and the second biopsy sample after warm ischemia and reperfusion (Fig. 1A). Donor and patient characteristics, as well as ischemia and reperfusion times, are displayed in Table 1. Consistent with previous studies in murine models of renal ischemia, we observed that human ENT1 and ENT2 transcript levels are repressed following warm ischemia and reperfusion (Fig. 1B). Hepatic protein levels of ENT2 are very low during ischemia and after reperfusion, whereas ENT1 protein levels show a stronger expression during ischemia and show a severe decrease following liver ischemia and reperfusion (Fig. 1C). We correlated the amount of ENT1/ENT2 protein expression to outcome parameters (e.g., aspartate aminotransferase [AST], alanine aminotransferase [ALT]), but based on the low number of biopsy samples we cannot state a correlation between the recovery phase of the recipient related to the amount of ENT protein expression in the liver biopsies. However, the expression levels of ENT1 and ENT2 were consistent with studies of murine Ent1 and Ent2 expression in a model of partial hepatic ischemia and reperfusion (Fig. 2A). Indeed, murine Ent1 and Ent2 transcript and protein levels were repressed following 45 minutes of liver ischemia and 2 hours reperfusion (Fig. 2B,C). Together, these studies demonstrate that hepatic ENT1 and ENT2 transcript levels are repressed during conditions of limited oxygen availability, indicating the likelihood of a transcriptional regulated endogenous protective pathway directed towards enhancing extracellular adenosine levels and signaling during liver ischemia and reperfusion injury.

Figure 2.

ENT1 and ENT2 expression in murine livers following hepatic ischemia and reperfusion. (A) The left lobe of the liver was exposed to 45 minutes of ischemia followed by 2 hours of reperfusion. (B) Ent1 and Ent2 transcript levels in left livers exposed to sham operation (0) or 45 minutes of ischemia followed by 2 hours reperfusion (n = 3-4 independent experiments). (C) Ent1 and Ent2 protein levels (β-actin to control for loading conditions; one representative blot of three is shown).

Pharmacologic Inhibition of ENTs With Dipyridamole Is Associated With Elevated Hepatic Adenosine Levels and Confers Liver Protection

After having shown that ENT1 and ENT2 are transcriptionally regulated during liver ischemia and reperfusion, as occurs during human liver transplantation, we next pursued studies to address their functional contributions to the regulation of extracellular adenosine levels and outcomes of hepatic ischemia and reperfusion injury. For this purpose, we exposed mice to 45 minutes of partial hepatic ischemia and 2 hours of reperfusion. In order to address the functional role of ENTs, we pretreated the experimental animals with intravenous dipyridamole (0.5 mg/25g mouse intravenously) 15 minutes prior to the onset of liver ischemia (Fig. 3A). Indeed, studies using high-performance liquid chromatography (HPLC) to measure hepatic adenosine levels in ischemic livers that were shock-frozen immediately following 45 minutes of hepatic triad occlusion revealed elevations of adenosine levels following ischemia. Importantly, these elevations of adenosine were further enhanced in mice pretreated with dipyridamole (Fig. 3B). Subsequent functional studies of the clinical outcome of hepatic ischemia and reperfusion injury revealed that mice pretreated with dipyridamole experienced attenuated plasma levels of AST and ALT and a less severe degree of hepatic tissue injury 2 hours (Fig. 3C,D) and 24 hours (Fig. 3E,F) following hepatic ischemia, indicating a protective role of dipyridamole in liver ischemia and reperfusion. Together, these studies demonstrate that ENT inhibition with dipyridamole is associated with liver protection from ischemia and reperfusion injury by way of enhancing hepatic adenosine levels and signaling events.

Figure 3.

Ischemic liver injury following inhibition of adenosine transporters during hepatic ischemia and reperfusion. (A) Experimental set-up to study liver ischemia and reperfusion with (black bar) and without (white bar) dipyridamole (DIP) treatment (0.25 mg/25g mouse intravenously). (B) Liver adenosine content measured immediately following 45 minutes of liver ischemia with or without DIP treatment. Wild-type mice with or without DIP treatment were exposed to (C,D) 45 minutes of liver ischemia and 2 hours of reperfusion or (E,F) prolonged reperfusion (24 hours) before liver function was assessed by measurement of AST, ALT, liver histology (200-fold magnification; one of three representative slides is shown) and quantification of histologic tissue injury (Suzuki Scoring Index, 0-4) (n = 4-6).

Selective Genetic Deletion of Ent1 Mediates Hepatic Adenosine Elevations and Liver Protection From Ischemia and Reperfusion Injury

After having shown that nonspecific inhibition of ENTs with dipyridamole is associated with elevated hepatic adenosine levels and concomitant liver protection from ischemia, we next pursued studies to address the functional contributions of Ent1 versus Ent2. For this purpose we exposed previously described mice for Ent1 or Ent2 to liver ischemia,[13, 19] and measured hepatic adenosine levels and assessed liver injury. Indeed, we observed that Ent1−/− mice experienced higher postischemic adenosine levels as compared to littermate controls matched age, gender, and sex when exposed to 45 minutes of partial liver ischemia (Fig. 4A). Moreover, Ent1−/− mice experienced less pronounced elevations of plasma ALT and AST (Fig. 4B) and histologic liver injury (Fig. 4C) following 45 minutes of liver ischemia and 2 hours of reperfusion. In addition, we observed that liver inflammation induced by ischemia and reperfusion was significantly attenuated in Ent1−/− mice (Fig. 4D). Moreover, second organ injury of the lungs induced by liver ischemia and reperfusion was significantly attenuated in Ent1−/− mice (Fig. 4E). In addition, we performed experiments with prolonged reperfusion times. In these experiments, we followed 45 minutes of liver ischemia with 24 hours of reperfusion. Indeed, Ent1−/− mice exhibited significantly lower levels of tissue injury as examined by elevations of the transaminases AST and ALT and liver histology (Fig. 4F,G) after 24 hours of reperfusion time. In contrast, Ent2−/− mice exposed to liver ischemia failed to demonstrate more pronounced elevations of ischemia-induced adenosine levels (Fig. 5A), and showed similar levels of liver injury and liver inflammation as corresponding littermate control mice (Fig. 5B-D). Moreover, secondary organ injury of the lungs was similar in Ent2−/− mice or controls following hepatic ischemia and reperfusion (Fig. 5E). In addition, liver injury was similar also after prolonged reperfusion time (24 hours, Fig. 5F,G). Taken together, these findings demonstrate for the first time a selective role for Ent1 in liver protection from ischemia and reperfusion injury.

Figure 4.

Ischemic liver injury in mice gene-targeted for ENT1. (A) Liver adenosine content measured immediately following 45 minutes of liver ischemia in Ent1−/− mice (white bar) or littermate controls (black bar). (B-E) Ent1−/− mice or littermate controls were exposed to 45 minutes of liver ischemia and 2 hours of reperfusion before liver function was assessed by measurement of (B) AST, ALT, (C) liver histology (200-fold magnification; one of three representative slides is shown) and quantification of histologic tissue injury (Suzuki Scoring Index, 0-4) (n = 4-6); (D) liver inflammation was assessed by measurement of IFN-γ, IL-6 protein levels, and neutrophil marker MPO in livers and (E) lung inflammation was assessed by measuring IFN-γ, IL-6, and MPO in lungs. (F,G) Ent1−/− mice or littermate controls were exposed to 45 minutes of liver ischemia and 24 hours of reperfusion before liver function was assessed by measurement of AST, ALT, liver histology (200-fold magnification; one of three representative slides is shown) and quantification of histologic tissue injury (Suzuki Scoring Index, 0-4) (n = 4-6).

Figure 5.

Ischemic liver injury in mice gene-targeted for the ENT2. (A) Liver adenosine content measured immediately following 45 minutes of liver ischemia in Ent2−/− mice (white bar) or littermate controls (black bar). (B-E) Ent2−/− mice or littermate controls were exposed to 45 minutes of liver ischemia and 2 hours of reperfusion before liver function was assessed by measurement of (B) AST, ALT, (C) liver histology (200-fold magnification; one of three representative slides is shown) and quantification of histologic tissue injury (Suzuki Scoring Index, 0-4) (n = 4-6); (D) liver inflammation was assessed by measurement of IFN-γ, IL-6 protein levels, and neutrophil marker MPO in livers and (E) lung inflammation was assessed by measuring IFN-γ, IL-6, and MPO in lungs. (F,G) Ent2−/− mice or littermate controls were exposed to 45 minutes of liver ischemia and 24 hours of reperfusion before liver function was assessed by measurement of AST, ALT, liver histology (200-fold magnification; one of three representative slides is shown) and quantification of histologic tissue injury (Suzuki Scoring Index, 0-4) (n = 4-6).

ENT1-Dependent Liver Protection From Ischemia and Reperfusion Injury Involves Signaling Events Through Adora2b

After having shown that pharmacologic inhibition or genetic deletion of Ents is associated with elevated hepatic adenosine levels, and concurrent protection from ischemic hepatic injury, we next pursued the hypothesis that Ent-dependent liver protection involves adenosine signaling. To address this hypothesis, we treated Ent1 gene-targeted mice with an Adora2a or Adora2b antagonist and thus examined if blockade of one of these adenosine receptors abolishes the protective effect of Ent1-dependent adenosine generation. The dosing for the antagonists were chosen based on previous publications showing an effect in organ injury.[23-25] While mice with pretreatment with the Adora2a-specific antagonist ZM241385 (2 mg/kg intravenously) showed a similar degree of liver protection as Ent1−/− without treatment (Fig. 6A,B), Ent1−/− mice with pretreatment with the Adora2b-specific antagonist PSB1115 (0.5 mg/25g mouse intravenously) were not protected compared to Ent1−/− without treatment (Fig. 6C-F). Together, these studies indicate that kidney protection mediated by the ENT inhibitor dipyridamole involves signaling events through Adora2b during ischemic hepatic injury.

Figure 6.

Role of adenosine signaling in Ent1−/− mice. (A,B) Ent1−/− mice were treated with an Adora2a-specific antagonist ZM241385 (2 mg/kg intravenously, white bar) or vehicle (black bar) and subsequently exposed to 45 minutes of liver ischemia and 24 hours of reperfusion before liver function was assessed by measurement of AST, ALT, liver histology (200-fold magnification; one of three representative slides is shown) and quantification of histologic tissue injury (Suzuki Scoring Index, 0-4). Ent1−/− mice were treated with an Adora2b-specific antagonist PSB1115 (0.5 mg/25g mouse intravenously, white bar) or vehicle (black bar) and subsequently exposed to (C,D) 45 minutes of liver ischemia and 2 hours of reperfusion or (E,F) prolonged reperfusion (24 hours) before liver function was assessed by measurement of AST, ALT, liver histology (200-fold magnification; one of three representative slides is shown) and quantification of histologic tissue injury (Suzuki Scoring Index, 0-4).

HIF1α-Dependent Regulation of ENTs and Adora Receptors Following Liver Ischemia and Reperfusion Injury

After having shown the impact of adenosine reuptake by way of ENT1 and adenosine signaling by way of the Adora2b receptor during liver ischemia and reperfusion injury, we next wanted to investigate the transcriptional regulated pathway of these proteins. Since previous studies identified HIF1α regulating ENT1 and Adora2b receptor expression we used a novel mouse line with deletion of HIF1α in hepatocytes (HIF1αloxP/loxP Albumin Cre+, Fig. 7A) and studied ENT1/ENT2 and adenosine receptor expression with and without liver ischemia. Interestingly, ENT1 and ENT2 transcript levels were at baseline higher in the conditional HIF knockout mice compared to the appropriate controls (Fig. 7B). Furthermore, neither ENT1 nor ENT2 were repressed following liver ischemia in contrast to the control mice. Moreover, the increase in Adora2b receptor transcript following liver ischemia in control mice was absent in HIF1αloxP/loxP Albumin Cre+ mice (Fig. 7C). These findings are consistent with previous studies that identified a transcriptionally regulated pathway for ENT1, ENT2, and Adora2b involving HIF.[15, 26] Together, these studies indicate that ENT1 and Adora2b are transcriptionally regulated by way of HIF1α during liver ischemia and reperfusion injury.

Figure 7.

Role of HIF1α in regulating hepatic ENT and Adora receptor expression. (A) HIF1α expression in isolated hepatocytes of Adora2bloxP/loxP Albumin Cre+ and Albumin Cre+ control mice with and without liver ischemia. (B) Ent1 and Ent2 transcript levels in left livers of Adora2bloxP/loxP Albumin Cre+ and Albumin Cre+ control mice exposed to sham operation (-I) or 45 minutes of ischemia (+I) followed by 2 hours reperfusion (n = 3-4 independent experiments). (C) Adora1, Adora2a, Adora2b, and Adora3 transcript levels in left livers of Adora2bloxP/loxP Albumin Cre+ and Albumin Cre+ control mice exposed to Sham operation (-I) or 45 minutes of ischemia (+I) followed by 2 hours reperfusion (n = 3-4 independent experiments).

Discussion

Hepatic ischemia and reperfusion injury significantly contributes to the mortality and morbidity of major hepatic surgery and liver transplantation. Moreover, therapeutic approaches to dampen ischemia and reperfusion-mediated tissue injury are extremely limited, and studies trying to identify novel therapeutic targets is an area of intense research. Based on previous studies showing that levels of the antiinflammatory signaling molecule adenosine are tightly regulated by adenosine transporters (particularly ENTs), we pursued the hypothesis that ENTs can be targeted to increase hepatic adenosine signaling and thereby mediate liver protection from ischemia and reperfusion. Indeed, these studies demonstrated that ENT1 is particularly expressed in the human liver, and ENT1/2 transcript levels are repressed following liver transplantation in humans. Functional studies with the ENT inhibitor dipyridamole demonstrated liver protection in conjunction with elevations of extracellular adenosine levels. Moreover, we observed a selective phenotype in Ent1−/− mice characterized by elevation of hepatic adenosine levels and profound hepatoprotection from ischemia and reperfusion injury. Subsequent studies with pharmacologic blockers of adenosine signaling revealed that the observed protection in Ent1−/− mice predominantly involves Adora2b. Furthermore, we could show that Ent1/Ent2 and Adora2b are transcriptionally regulated by way of HIF1α by utilizing conditional mice. Taken together, these studies demonstrate a functional role for ENT1 in liver protection from ischemia and reperfusion injury and implicate ENT inhibitors in the treatment of ischemic liver injury.

The present findings demonstrate attenuated ENT1 and ENT2 transcript levels following ischemia and reperfusion during human liver transplantation, or during murine liver ischemia and reperfusion. These studies are consistent with previous findings showing repression of ENTs during conditions of limited oxygen availability. In fact, ambient hypoxia exposure of vascular endothelia,[12] cardiac myocytes,[27] or intestinal epithelial cells[12, 15] is associated with repression of ENT1 and ENT2 transcript and protein levels. Studies on the regulatory mechanism coordinating these responses revealed that both the ENT1 and the ENT2 promoter contain binding sites for the transcription factor HIF.[12, 15] Subsequent studies with transcription factor binding assays, promoter constructs, or HIF loss- or gain-of-function revealed that HIF directly binds to the promoter regions of ENT1 or ENT2, and mediates ENT repression during hypoxia. We could confirm these findings by using a transgenic mouse line with a floxed HIF1α gene to generate a mouse line with deletion of HIF1α in hepatocytes. The repression of hepatic ENT1/ENT2 following liver ischemia was absent in these mice. Furthermore, the induction of Adora2b receptor following liver ischemia was abolished, indicating that these proteins are transcriptionally regulated by way of HIF1α. Indeed, HIF is responsible for the transcriptional regulation of a coordinated response that results in increased extracellular adenosine signaling effects during hypoxia. In addition to repression of ENT1/ENT2, this response includes the transcriptional induction of CD73, the key enzyme for extracellular adenosine generation,[24, 28-32] and the Adora2b receptor.[24, 33-37] In addition to transcriptional repression by direct binding of transcription factors to a gene promoter, transcriptional repression is frequently mediated by transcriptional induction of microRNAs (miRNAs). Previous studies had shown that ENT1 or ENT2 are regulated during conditions of ambient hypoxia by direct binding of HIF1α to the promoter of ENT1 or ENT2, respectively.[15, 26] However, it is also conceivable that ENT repression could be mediated by HIF-dependent induction of miRNAs that would target ENT mRNA. Indeed, several previous studies have implicated miRNA induction and subsequent transcriptional repression of target genes during conditions of ischemia or hypoxia.[2] Several previous studies have demonstrated a protective role of adenosine signaling during inflammatory conditions. Indeed, the first report that pathophysiologically induced extracellular adenosine signaling by way of the Adora2a receptor is critically important and nonredundantly responsible for the immunosuppression during inflammation in vivo in the absence of any drug comes from a landmark paper from the research group of Dr. Sitkovsky.[16, 38, 39] Subsequent in vivo studies from the laboratory of Dr. Ravid suggested that also signaling events through Adora2b can dampen vascular inflammatory responses in response to endogenous elevations of extracellular adenosine levels in vivo.[40] Moreover, pharmacologic studies from Dr. Bruce Cronstein's laboratory established a functional role of adenosine receptor signaling in attenuating inflammatory cell activation.[41] Moreover, studies from the laboratory of Dr. Joel Linden demonstrated that activation of Adora2a receptors on inflammatory cells—particularly on natural killer T-cells—are involved in liver protection from ischemia.[10] In contrast to these studies, the present findings implicate Adora2b in ENT-mediated liver protection from ischemia. Consistent with these findings, several previous studies had implicated Adora2b in tissue protection from ischemia.[24, 35-37, 42-45] In addition, it is conceivable that the timing of the injury model may contribute to such differences; while early on (e.g., 2 hours after reperfusion) the dominant protective pathway could involve Adora2b, later inflammatory changes (particularly involving T-cells) could be attenuated by Adora2a.

Several studies have demonstrated that while adenosine signaling through Adora2b may be beneficial in an acute setting, this adenosine protection can become detrimental when it is prolonged.[46-49] Indeed, studies in a chronic liver disease model have shown detrimental effects of Adora2b signaling, using fatty liver disease—commonly associated with alcohol ingestion and abuse—as a model.[50, 51] During ethanol metabolism, adenosine is generated by the enzyme ecto-5′-nucleotidase, and adenosine production and adenosine receptor activation are known to play critical roles in the development of hepatic fibrosis. Dr. Cronstein's laboratory team therefore investigated whether adenosine and its receptors play a role in the development of alcohol-induced fatty liver. Wild-type mice fed ethanol on the Lieber-DeCarli diet developed hepatic steatosis, including increased hepatic triglyceride content, while mice lacking the ecto-5′-nucleotidase CD73 or Adora1 or Adora2b receptors were protected from developing fatty liver disease. These studies indicate that adenosine generated by ethanol metabolism plays an important role in ethanol-induced hepatic steatosis by way of both Adora1 and Adora2b and suggest that targeting adenosine receptors may be effective in the prevention of alcohol-induced fatty liver.[50]

Hepatic ischemia and reperfusion injury significantly contributes to morbidity and mortality of surgical patients undergoing liver transplantation. Indeed, the present studies reveal several lines of potential treatment modalities that could be used to prevent or treat hepatic ischemia and reperfusion injury. As a first line of treatment, the present studies suggest that HIF activators could be used to treat liver ischemia and reperfusion injury. Such compounds would result in repression of ENTs, thereby promoting adenosine-dependent liver protection. At the same time, these compounds would also increase extracellular adenosine production and signaling, by transcriptionally inducing enzymes that produce adenosine during ischemic conditions.[1-4] Interestingly, a recent clinical trial shows that HIF activators can be safely used in patients for the treatment of renal anemia.[5] A second line of treatment could be to use the ENT dipyridamole—either as a means of prophylactically treating patients during liver surgery, or as an addition to liver conservation solutions during liver transplantations. Again, dipyridamole has a great safety record in patients, for example, for the prevention of recurrent stroke or to maintain patency of dialysis grafts.[6, 7] As a third approach, adenosine receptor agonists—particularly for Adora2b could be used. In fact, we recently described and characterized a highly selective adenosine receptor agonist, BAY 60-6583. Finally, additional therapeutic approaches that would enhance hepatic conversion of ATP to adenosine, for example, by treating with soluble apyrase (conversion of ATP/ADP to AMP),[8-11] or nucleotidase (conversion of AMP to adenosine)[12, 13] could be considered.

Taken together, the present studies provide evidence that ENT1 (and to a lesser degree ENT2) is expressed in human livers. Subsequent studies in mouse models of liver ischemia and reperfusion point towards a therapeutic role of ENT1 inhibition in this model, as it is associated with elevated hepatic adenosine levels and protective signaling effects through the Adora2b receptor. Future challenges will include clinical studies with ENT inhibitors or Adora2b agonists to examine if the present findings can be translated from bench to bedside.

Acknowledgment

The present research work was supported by 1 KO8HL103900-01 to MZ, an American Heart Association Grant to AG and National Heart Institute Grants R01 DK097075, R01-HL0921, R01-DK083385, R01-HL098294, POIHL114457-01 and a grant by the Crohn's and Colitis Foundation of America (CCFA) to HKE.

Authors Contributions

M.Z., A.G., E.T., S.E., M.K., A.G., M.R.B. researched data. D.S.C., I.K. provided new research tools. H.K.E., A.G., M.Z. wrote the article.

Ancillary